8/2/2019 Cowhey Aronson Richards
1/21
PART II: POLITICS
Shaping the Architecture of the U.S. Information andCommunication Technology Architecture: A PoliticalEconomic Analysis1
Peter F. CowheyUCSD, Graduate School of International Relations and Pacific Studies
Jonathan D. AronsonAnnenberg School for Communication and School of International Relations
John RichardsUCSD, Global Information Industry Center
AbstractHow did political economy help shape the revolution in telecommunications and computer networking?We offer three arguments concerning the impact of political economy and policy on the architecture ofthe U.S. information and communication technology infrastructure. First, it tilted toward an archi-tectural principle of modularity that influenced the paths of both the telecom equipment, computerequipment and software, and computer networking markets. Second, it created multiple network infra-structures for telecommunications when other countries either tried to retain a monopoly infrastructureor limit the number of competitors. Third, it propelled a particular architecture for computing (intel-ligence at the edge of the network) and the full realization of the potential benefits of the Internet. Theparticular policy mix for competition matters, and this policy mix reflects fundamentals of politicaleconomy.
KEY WORDS: computing intelligence, decentralization, electoral institutions, federalism, technological
trajectories
This article analyzes how policy and politics shaped the architecture of the globalinformation and communication technology infrastructure, the ICT infrastructure.
This intersection of computing with the telecommunications network, which now is
embodied in the Internet and the Web, changed dramatically from the mid-1950s
to the end of 2000.
The potential for a radical change in the ICT infrastructure grew from thecapabilities in computing, electronic switching, and fiber optics that sprang from the
microelectronics revolution. The United States made the first and most radical break
with monopoly in telecommunications and led the revolution in computer network-
ing. But political economy factors shaped the specific technology path for the market
place. Two key questions are asked. First, why did the United States take the lead in
redefining market structure away from a telecom monopoly and toward competition
based on a competitive long-distance infrastructure? Second, how did Americas
particular policy mix influence the architecture of the ICT infrastructure?
These two questions share a common premisetechnological trajectories
reflect policy and market conditions.2 Specifically, U.S. policy influenced the tra-
jectory in three ways. First, it tilted toward an architectural policy principle of
modularity, whose implementing policy norms influenced the paths of both the
105
Review of Policy Research, Volume 26, Numbers 12 (2009) 2009 by The Policy Studies Organization. All rights reserved.
8/2/2019 Cowhey Aronson Richards
2/21
telecom equipment and computer networking markets.3 At an intuitive level,
modularity works in the same way as Lego building blocks. Many pieces of
different shapes can be mixed and matched to form different forms because they
have standardized interfaces that allow them to interconnect.4 Second, when
other countries tried to retain a monopoly infrastructure or to carefully limit
the number of competitors, it created multiple network infrastructures for tele-communications. Third, U.S. policy enabled a specific computing architecture
(intelligence at the edge of the network) and the realization of the potential ben-
efits of the Internet. The process involved much more than the United States
embracing competition in networking traditional telecom and new information
services.
These questions frame the issue concerning the emergence of competition in the
United States rather differently than most economic analyses. Most economists
focus on the degree to which the United States shifted in the direction of ideal
competitive markets. Seen this way, the U.S. policy shifts arguably had deep flaws.
5
But this is the wrong approach; it is too narrow. The remarkable part of the U.S.
story was how its trailblazing move toward a certain type of competition set the stage
for the successful emergence of the Internet and the Web.6 The reasons that
the United States chose its particular policy mix are revealing lessons in how the
contours of a countrys political architecture influences its policy propensities. The
policy predilections in turn influence technology markets. From the Olympian
heights of microeconomics, the blend of policies was imperfect, but that is not the
key issue.
This case study lays out the policy and technology developments that shaped
the American ICT infrastructure. In the mid-1950s, the ICT infrastructure beganto emerge from the joining of two distinct marketsa monopoly telecom mar-
ketplace (the old AT&T monopoly) and a distinct, concentrated computer and
software industry centered on mainframes (dominated by IBM), and then mini-
computers (such as DEC). During the 1960s and 1970s, the growth of computer
networking and nascent competition in terminal equipment (whether a phone or
a computer) tied to the telecom network created a new market for value-added
information services that regulators permitted as a limited exception to the
telecom monopoly.7 As data networking matured, an integrated ICT market
gradually emerged. The tensions over the commercial model for the merging
ICT infrastructure helped to drive the breakup of AT&T in 1984. Later, a com-petitive telecommunications market bolstered a new model of computing that
relied on the personal computer tied to office and wide area networks. This
model came to full blossom in the form of the Internet and the Web. It ultimately
reestablished U.S. leadership in the global ICT industry that was then under
serious challenge from Japan.
We rely on a case study to show why the specific policy mix chosen for ICT
conforms to a well-established framework for understanding American political
economy. But it emphasizes the consequences of particular political and policy
compromises for the performance and architecture of the ICT infrastructure.
Political economy means that policy rarely emerges in a pristine textbook manner.
The twists of policy compromise influenced much of the architectural nuance of
technological change.
106 Peter F. Cowhey, Jonathan D. Aronson, and John Richards
8/2/2019 Cowhey Aronson Richards
3/21
At the same time, note what is not claimed. Although this path of innovation
ultimately paid high returns to the United States and refurbished its global
competitive leadership, there is nothing in this study to determine which national
political economic structure is best for innovation or economic performance.8
The first part explains why the United States moved first on competition. We
argue that the answer is rooted in Americas political economy, but we divergesubstantially from explanations that focus mainly on the role of American courts or
on interest-group politics. We point to the consequences of American institutional
structureespecially divided government, federalism, and winner-takes-all federal
electionsinteracting with a particular constellation of interest-group coalitions
(the corporate competition coalition). Our conclusion concerning the conse-
quences of policy, discussed in the second and third part of the article, explains the
specific impacts on the ICT infrastructure. To simplify the exposition, we concen-
trate on the political economy of the telecom side of the ICT infrastructure and
its consequences for the architecture of the infrastructure. Major changes in theinformation technology industry are noted throughout the analysis without delving
into the same level of detail.
The Political Economy of the United States
At the core of the argument concerning the political economy of markets are
political institutions and their impact on the incentives and authority of elected
politicians to shape marketplace outcomes to the advantage of specific sets of
constituents. This section sketches the institutional argument with examples from
the ICT eras policy in the United States. The second and third parts of the articleexpand on these illustrations.
The U.S. political system has three features that are salient to the setting of ICT
policythe division of powers, the majoritarian electoral system, and federalism.9
First, the division of powers in the U.S. government was designed to make it
difficult to initiate major policy changes, but also difficult to rapidly undo major
commitments. The division between the president and Congress (and the division
between the two chambers of Congress, one based on population and the other
based on equal representation for the individual states) creates many veto points
during the decision process where an initiative can be stopped. 10 This hampers the
passage of major changes in laws that have sweeping geographic consequences anda wide range of winners and losers. Thus, only two comprehensive U.S. telecom-
munications laws passed during the twentieth century, the 1934 and 1996 congres-
sional acts.
These two Acts delegated to the Federal Communications Commission (FCC)
much of the decision making concerning federal communications and computer
networking policy. The FCC shares jurisdiction on competition policy with the
Antitrust Division of the Department of Justice (which largely focuses on specific
complaints concerning violations of competition laws).11
The inherent conflict between the executive and legislative branches means that
Congress is less willing to grant the kinds of discretion to executive bureaucracies
that are found in parliamentary democracies, where the division between the
executive and the legislature is blurred. Congress does, however, recognize the
Shaping the U.S. ICT Architecture 107
8/2/2019 Cowhey Aronson Richards
4/21
need for a substantial amount of expert bureaucratic authority. Thus, the FCC is
allowed to deal with complicated issues where many of the detailed political costs
and benefits are difficult to determine. Congress then uses a variety of specialized
controls to shape the power delegated to the bureaucracy.
In the case of the FCC, Congress confirms the commissioners nominated by the
President and stipulates the division of political representation of commissioners.It mandates a 32 split, with the Presidents party holding the majority and
the chairmanship. This replicates the political sensibilities of the major parties
within the majority voting system of the Commission. Commissioners introduce an
element of political sensibility into the Commissions work, but normally have a
narrower range of ideological divisions and a more technocratic slant than Con-
gress. Congress also uses the power of the purse by threatening to use its budgetary
powers to instruct the FCC on certain matters. For example, Congress directed the
FCC not to use public funds to create rules to auction communications satellite
spectrum. Similarly, the Congress requires transparency in decision making so thatall of the key interests will have access to the decision process. This allows congres-
sional members to observe the process with an eye to its politics, and to pressure the
Commission if there is a compelling political interest.
The net result is an FCC that is responsive to presidential and congressional
politics, but legally empowered to make important discretionary policy. It is subject
to judicial review for its adherence to a reasonable reading of the underlying law. 12
It takes decisions based on its analytic judgment, the evidence on the public record
developed in each procedure, and an instruction to use this discretion to serve the
public interest. These expert and transparent, but politically informed, decisions
shape market dynamics.A second feature of the U.S. political institutions matters for shaping policy.
The presidential and congressional election system is based on winner-take-all
voting. Analysts of electoral systems have shown that this voting system builds
a strong interest in brand identity for political parties.13 Despite the role of
lobbying and campaign contributions, parties invest to develop a reputation with
policy initiatives on broad issues that they think will mobilize elite and mass
electoral support.
Naturally, ICT infrastructure policy mattered to the high-technology industry
and research communities. It also achieved broad political salience to the voting
public in two ways. First, as a matter of equity, there was continuing sensitivity totelephone pricing and service accessibility, and later broadband pricing. Second,
infrastructure policy was part of the broader debates over economic policy, includ-
ing the debates over deregulating the American economy and the creation of
the new or Internet economy to stimulate growth. For example, the Clinton
administration highlighted its telecommunications policy to polish its reputation as
procompetition and pro-innovation Democrats.14 It bragged about early U.S. lead-
ership in adopting the Internet. Similarly, the Bush administration later worried
about the potential embarrassment of Americas lagging position on deployment of
broadband.
The third feature of the institutional context is federalism, the division of author-
ity between the federal and state governments. The U.S. Constitution reserves all
powers for the states that are not given explicitly to the federal government. Each
108 Peter F. Cowhey, Jonathan D. Aronson, and John Richards
8/2/2019 Cowhey Aronson Richards
5/21
state is allocated two Senators, regardless of its population. This increases the power
of low-population farm and mining states at the expense of the large, urban-
centered states.
Federalism matters for ICT policy directly and indirectly. It has a direct
impact, because the subsidy of rural communications users and providers is a
powerful constraint on all regulatory policies influencing pricing and competition.This spilled over from traditional phone services to computer networking in the
1990s as the Internet became a mass medium. It indirectly matters because fed-
eralism also provides a foundation for strong competition policy. State authorities
often have used competition policy to shelter local competitors from nationally
dominant competitors, and these local firms also enlisted the support of their
senators.
The key role of rural states in the Senate also heightened interest in competition
rules that emphasized consumer welfare over the welfare of competing firms,
because those states have less interest in industrial policy favoring national cham-pions.15 The cumulative result was an American economy with broad geographic
scope for its competitive firms and far less concentration in its key industries,
including telecommunications and electronics, than its counterparts in other major
countries.16 It also had a telecom market whose behavior was skewed by a pricing
structure that bore little relationship to efficient costing. The implications for ICT
infrastructure policy and architecture were profound.
The Political Economy of Value-Added Competition: 19501983
The first phase of convergence of computing, software, and communications beganin the mid-1950s and extended through 1983. The existing telecom infrastructure
was not well suited for the growth of computer networking. Transmission quality
and throughput were optimized for voice networks, which made adding data
networking capabilities difficult and expensive. Until the mid-1970s, network intel-
ligence (computing power to guide connections and traffic routing on the network)
was expensive and highly centralized in a set of switches at the top of the pyramid
of less capable switches routing local and regional traffic.17 Network transmission
capacity on pairs of twisted copper wires was also sparse, expensive, and specialized.
It was expensive and physically difficult to expand capacity.
Early computer networking services were slow and geared toward large businessusers. Quality voice services required 64 kbps, but data rates on these circuits were
slower and less reliable. When computer networking took off, issues involving the
quality, the speed of transmission, and related technical issues made the traditional
networks practices inadequate for the new data networks.18 At the same time, the
post-1945 era featured the growth of population and economic activity in the South
and the West, thereby generating more demand for continental-wide networks and
the growth of U.S. multinational firms. This put further pressure on an infrastruc-
ture that had been created and priced with an eye toward local and regional
activities.19
Even with a national governing system prone to deadlock, policy evolution can
move quickly if economic interests and political institutions are aligned, as was
the case during the first two ICT eras. Through the 1970s, the United States was
Shaping the U.S. ICT Architecture 109
8/2/2019 Cowhey Aronson Richards
6/21
economically distinct on the global scene, because it was by far the largest ICT
market and its economy had a continental scope. Policies could stimulate new
entrants while simultaneously retaining market scale and scope for incumbents.
In addition, large enterprise users were deploying ICT as a central tool to enable
new production and service processes to stay competitive nationally and
internationallyairline reservation systems were pioneers in computer network-ing.20 Large enterprises also began using long-distance telephony intensively to
increase branch coordination.21 This produced a group of large potential customers
concerned with the markets organization among sophisticated firms. This in turn
guaranteed an environment favorable for political entrepreneurship.
The role of large users bears special attention, because in the postwar era, ICT
gradually transformed them, thus intensifying their policy advocacy over time. ICT
became more significant then as a major cost factor for them. The United States was
the worlds leader in multinational companies. In response to rising competition
over time, its financial institutions and many manufacturing firms have becomeinformation analysis companies, which deliver this information in the form of, for
example, global financial or engineering products. Global banks no longer focus
primarily on checking or mortgages. At least until the financial meltdown of 2008,
their edge came from complex financial products that rest on high levels of com-
puting and global information operations that are rolled out quickly on a global
scale over their ICT infrastructures.22 Multinational manufacturers know that the
cost and quality of production are important, but the information-intensive, global
design, and service functions are their critical edge. By the late 1980s, Boeing
executives sometimes joked that an airplane is a flying software stack because there
was more value added in the complex programming than in the sophisticatedairframe.23 This fundamental shift in the strategic use of ICT drove these firms to
become committed advocates for changes in ICT markets.
The political institutional legacy of the American market structure shaped the
way that these emerging interests played out. No firm legal basis for the national
AT&T long-distance monopoly existed, and many smaller telecom carriers
remained in small states and rural areas.24 This lacuna arose because it always
proved politically difficult to craft legislation to authorize, or, later, to preserve a
monopoly. In addition, a diverse industry of large and small electronics firms had
grown to serve consumer, industrial, and defense markets. They lusted to supply
equipment to American telecommunications networks. In 1956, their continuingcomplaints against AT&T benefiting from research subsidized by monopoly utility
rates led to a limited antitrust victory that forced Bell Labs to license its technology
to them at little or no cost. Also in 1956, the first limited liberalization of terminal
equipment attached to the network was mandated. Then, in 1968, the FCC,
in its Carterfone decision, started down the path to full freedom of competition
in equipment attached to the network.
Meanwhile, federal power sharing with the states over pricing policy and a
Senate sympathetic to rural interests restricted the ability of AT&T to lower long-
distance prices. So long as AT&T made high margins on long-distance service, it
could then transfer part of the funds to support smaller rural carriers.25 Although
AT&T offered special discounts to large corporate customers, it could not offer true
cost-related pricing, so large customers continued to seek market change.
110 Peter F. Cowhey, Jonathan D. Aronson, and John Richards
8/2/2019 Cowhey Aronson Richards
7/21
The growth of computer networking, especially by IBMs smaller rivals, created
another powerful set of motivated allies that were unhappy with AT&Ts domi-
nance.26An IBM plug-compatible industry grew up, which targeted the network-
ing market. This led directly to the formation of a corporate competition coalition
comprised of computer companies that wanted to create customized computer
networks or feared AT&Ts entry into the computer equipment market. The com-puter companies were joined by large corporate clients, smaller electronics equip-
ment vendors, would-be resellers of basic phone services, and government agencies,
all seeking better deals.27
A new principle of modularity slowly emerged to guide governance.28 Modularity
recognized that service and equipment reliability and efficiency did not require
uniform design and supply of a system. Instead, the system could be designed to
allow a variety of subsystem designs and suppliers while still working together
smoothly as a whole by requiring transparent and nondiscriminatory interfaces
for components. This reasoning led to a reconceptualization of the market andthe foundations for its regulations. It became common to distinguish among basic
phone services provided over the general public network, the equipment that
enabled it, and new advanced communications and equipment functions made
possible by new electronic and computing technologies. Momentum grew to com-
petitively deploy new value-added services and equipment.
Liberalization of terminal equipment markets created the first policy norm
flowing from modularity that drove equipment architecture. The norm held that
new equipment attached to the network was acceptable if it did no harm to the network. The
FCC finally recognized that the demands for computer networking required less
restrictive equipment markets.29
This meant that equipment architecture had toallow for logically separate functionality with an open interface that allowed differ-
ent pieces of equipment from any vendor to be linked together. A closed interface
would not allow another vendors equipment to be linked.
Although only incipient at the time, the no harm to the network norm implied
a freedom of choice that grew into a second norm, technology neutrality. This norm
deeply resonated with U.S. political and market institutions. America rarely picks
civilian technology champions with any consistency.30 Its diverse economy does not
easily generate political agreement on a single technology path for any market. By
the 1980s, American policy makers also consciously doubted their ability to readjust
their direction if they chose the wrong technology path, and so neutrality seemeda policy of substantive and political merit. This norm subsequently shaped the
crucial choices made concerning the second generation of wireless technology
introduced in the early 1990s. Unlike most other countries, the FCC opted for a
policy of technology neutrality. The deployment of multiple wireless architectures
resulted. The cost of diverse architectures was some confusion and delay in deploy-
ment of features requiring mass scale, although over the long haul, it had benefits
for innovation that emerged in the late 1990s. This was similar to earlier computer
industry developments in the United States.31
At the same time, the Commission lurched toward allowing competition in the
provision of networked computer services. In 1959, AT&T Long Lines established
a discount rate for its largest corporate and government customers that was a
reasonable proxy for a wholesale price for the leasing of transmission capacity to the
Shaping the U.S. ICT Architecture 111
8/2/2019 Cowhey Aronson Richards
8/21
new data networks. The FCC embraced this benchmark when it forced AT&T to
lease capacity at a wholesale price to computer networks.32 In doing so, the FCC
embraced for narrow purposes what eventually became a third norm, implementing
modularitygovernment mandates requiring network operators with market power to share
their network capacities with competitors on terms that would not restrict competition.33
In the realm of monopoly economics, government pricing intervention iscomplex because of the difficulty the monopolist faces when trying to differentiate
prices among different classes of customers whose elasticity of demand varies. It is
also a political swamp because everybody makes special claims concerning rates.
The political goal was to show that pricing was friendly to household consumers,
especially in rural and low-income areas. This clashed with the logic of the net-
works cost structure. For example, costs were higher in rural areas where longer
transmission distances supported fewer customers. Moreover, given the large
common costs of networks, such as billing systems, attributing costs to different
services and areas involved creative, albeit government-dictated, accounting. Ingeneral, the pricing formulas caused denser urban areas to subsidize rural areas,
and long-distance customers to subsidize local service users.34
Monopoly had rested significantly on its promise of preserving stable prices
for local rates, and subsidized prices for rural and high-cost areas. A key political
litmus test for these competitive reforms was whether they could be reconciled
with these pricing arrangements. Over time, a quiet policy practice developed
that the deployment of value-added, competitively provided services should not
undermine the basic pricing and service structure of the general public network.
By linking prices for sharing the AT&T network to an established rate (the whole-
sale price for large customers that existed under monopoly), the FCC laid thebasis for skirting the political problems involving monopoly pricing. Anchoring
the AT&T price for leasing to computer networks to existing pricing for large
customers was politically reassuring because data services were added to a pre-
existing rate compromise that AT&T had promised would not upset consumer
pricing.
At the same time, side-stepping major pricing reform also opened the wedge for
allowing private networks to join together geographically far-flung firms offices
with capacity leased from AT&T at wholesale rates. MCI applied for permission to
provide specialized corporate services over its own microwave network in 1962, and
in 1969, won approval for its first link, between Chicago and St. Louis. When theFCC generalized this decision in 1971, only about 3 percent of the total Bell system
revenue was at stake.35 The FCC also allowed private line carriers to interconnect
with AT&T facilities. Predictably, the battle over the precise terms of interconnec-
tion led to MCI, and later to Justice Department suits, which led to the decision to
divest AT&T that took effect on January 1, 1984.36
Even as telecom changed, IT also was restructured. IT in this era was charac-
terized by deployments of computers of limited processing capacity clustered
in central offices, manufacturing plants, and research sites. A large share of the
software was supplied by the computer maker. Early on, governments and large
enterprise buyers were the major users of IT, and, later, networked IT. Just as in
telecom, a combination of larger and smaller rivals and some big IT customers
pressed for ways to curb IBMs early dominance of the market.
112 Peter F. Cowhey, Jonathan D. Aronson, and John Richards
8/2/2019 Cowhey Aronson Richards
9/21
In 1956, the IBM Plug and Play compatible antitrust decision partially separated
hardware from software and led to the birth of the minicomputer industry. During
the 1960s, further stresses to this industrial structure emerged as the growth of
computer networking allowed new entry and software strategies in the IT industry,
and integrated circuits became microprocessors. This allowed economies of scale in
the production of chips by independent suppliers, a strong boost for rivals to IBM. 37
Furthermore, experience with mainframe computing produced a growing pool of
programmers who could write code independent of the big computer companies.
Here, public policy was at the core of market evolution. The U.S. government
antitrust suit against IBM led to a 1969 agreement that caused IBM to decouple
hardware and software. This allowed the take-off of an independent software
industry that featured packaged software.38 Although still a far cry from the soft-
ware industry associated with the PC industry, this began to erode IBMs domi-
nance and added to the move toward the modularity principle in computing
hardware and software. Still, IT remained largely the domain of significant dataprocessing applications for the largest government and enterprise buyers.
In short, the first change in policy introduced the principle of modularity. The
policy norms of no harm to the network, technology neutrality in government
regulatory policies for ICT, and compulsory sharing of network capabilities by
carriers with market power emerged. This approach initially was applied to
create limited competition in value-added information services, private network
services, and competition in terminal equipment. The debate over the terms for
equipment competition, leasing of network capacity for the new computer networks
emerging in the 1970s, and private corporate services smoothed the entry of
the computer industry into the telecom policy realm. Computer vendors and theirlargest customers wanted to network expensive mainframe computers to allow
more efficient cost sharing and operations. Networking became a major feature in
struggles for future control of the IT industry, and all major technology suppliers
and large network users pushed for policy change on networking.39
The Political Economy of Competing Infrastructures: 19842000
The break up of AT&T and the introduction of competition in the long-distance
services and network facilities markets sparked a global revolution in the organiza-tion of the telecommunications industry. It made possible a further revolution
in computing and broadcasting. By doing more than just forcing networks with
market power to share their facilities on procompetitive terms, a second principle
emerged for policy. The reasoning was that regulating the shared use of a monopoly
infrastructure seemed complicated and unlikely to create innovations in infrastruc-
ture that might emerge from a network designed from scratch. At the same time, no
entrant was likely to roll out a national network quickly, thus diminishing the value
of network externalities (more connections make a network more valuable) for its
customers. Over time, a second organizing policy principle took form: encourage the
emergence of competing network infrastructures by removing legal barriers to their
creation and by forcing the dominant incumbent to share its network with its rivals
until competition in the market is robust.
Shaping the U.S. ICT Architecture 113
8/2/2019 Cowhey Aronson Richards
10/21
Why did change unfold as it did between 1984 and 2000? Why did change begin
in the United States? If competition was driven mostly by technological and market
forces, why did it unfold so differently in the United States than elsewhere?
During the late 1970s, the combination of slow economic growth and high
inflation in the United States pushed deregulation of public utilities onto the
presidential and congressional political agendas. Political parties strive to benational policy entrepreneurs. Democrats and Republicans both saw deregulation
as a way to show their commitment to revive the American economy.40 The political
economic interests of the corporate competition coalition reinforced their enthusi-
asm for deregulation.
Renewed antitrust action during the Carter administration led to the break up of
AT&T during the Reagan administration. The decision reflected American political
institutions. First, the courts followed established antitrust law favoring a consumer
welfare standard that arose from the U.S. political economy. Second, the president
and Congress cannot easily take decisive legislative action to steer an industrybecause it can be blocked at many points. This structural factor sidetracked AT&Ts
attempt to legislatively assert its monopoly to repulse growing pressure from MCI
and other upstarts and convinced a generation of entrepreneurial politicians that
identification with AT&Ts critics was politically advantageous. Congressional
Republicans and Democrats both were wary of protecting monopoly, as it became
a national political issue. Even the Democratic Party, predisposed to supporting
organized labor and therefore a likely ally of AT&T and its union members,
spawned a generation of Atari (computer) Democrats critical of monopoly. Eco-
nomic conservatives in the Republican Party joined them. This coalition sufficed to
block preemptive legislation to preserve the phone monopoly.41
Third, although many in the Reagan administration were wary of the AT&T
antitrust decision, the White House did not try to overturn it, because the staff saw
it as politically risky to favor monopoly.42 Fourth, the settlement made political
sense because it could withstand pressures to protect incumbents before and after
the AT&T breakup. The long-distance competition facing the new AT&T was
balanced by monopoly phone services for the new regional Bells. This settlement
protected both local and rural telephone service pricing. The FCC and state utility
commissions could mandate cross-subsidies from long-distance carriers to local
phone monopolies and still allow competition to improve services and lower long-
distance pricinga rough rule of thumb was that until the late 1990s, the cross-subsidy amounted to about 40 percent of the revenue from a long-distance call.
Competition yielded large efficiencies and pricing much closer to underlying
cost structures even with these cross-subsidies. As a result, long-distance prices
dropped steadily and significantly, a development that appealed to the middle
class that tended to vote more than other Americans. Network competition also
appealed to the corporate competition coalition because they received better
prices and faster service innovations from the new entrants, which took special aim
at the business market.43
The political compromise on rates produced numerous distortions in the market,
but it was the price of a policy change. It is instructive to see how telecom pricing
compared with the political and regulatory dynamics attached to the cable television
industry. Cable began as a series of locally granted monopoly franchises and quickly
114 Peter F. Cowhey, Jonathan D. Aronson, and John Richards
8/2/2019 Cowhey Aronson Richards
11/21
won legislative favor as a way of delivering television to rural and urban areas with
interference problems.44 In 1984, Congress passed a bipartisan Cable Act that
ended local price regulation and banned the Bells from purchasing cable systems.
The idea was that the still-young industry had an incentive to price attractively and
would invest more aggressively than the monopolistic Bells. The legislation was
especially attractive to Western members of Congress in both parties because ofvast rural areas and the clustering of corporate headquarters for cable in that region
of the country.
As cable boomed, it became a powerful industry with revenues far exceeding
the big three network broadcasters. In the late 1980s, as cable prices climbed rapidly
and service remained spotty, it became a grass-roots consumer issue that crossed
party lines. (This also gave the network broadcasters and the Bells a chance to strike
back at their cable rivals by allying with consumer groups.) The ClintonGore
campaign turned it into an issue to show that the White House was out of touch with
consumers. Two-thirds bipartisan majorities in the House and Senate passed theCable Rate Act of 1992 and overrode President Bushs veto (the first Bush veto to
be overturned). The Act instructed the FCC to work with local municipalities on
capping cable rates, a sharp reminder that consumer pricing in the communications
industry could prompt swift Congressional intervention.45
The outline of a regime that would dominate the United States emerged from
the struggle over the fate of AT&T. The principle of favoring competitive network
infrastructures built on the earlier norm of forcing dominant networksthose
controlling essential bottleneck facilities, such as local transmission networksto
share their capabilities with new rivals in order to promote rapid entry into the
industry. This required detailed FCC supervision of dominant carriers on inter-connection. The challenge was to police against bottlenecks in a way that allowed
market forces to rationalize costs, staffing, and prices. Tools such as price caps and
dominant carrier regulation were designed to foster procompetitive interconnec-
tion with new entrants and allow pricing rationalization.46
In addition to caution on pricing of local services, politics dictated two other
policy practices. To cash in on the political promise of competition, regulators used
competitive reforms to promote technological and service innovation for ICT, not
just lower prices. Economic theory had long set the standard of maximizing con-
sumer welfare. This practice clarified what political leaders meant by consumer
welfareit included technological innovation that was dear to the coalition sup-porting competition. In addition, policy makers were sensitive to employment
effects. For example, they could allow labor staffing in dominant incumbents to
decline, but needed to cushion job losses by encouraging the entry of new compa-
nies, which might offset the downsizing of old incumbents. Clearly, competition
would lower the staffing levels from the days of monopoly, but there was enormous
concern about the pace of adjustment.
This mixture seemed politically successful. Prices for long-distance and data
services dropped significantly. Service innovation climbed. Computer networking
continued, steadily progressed, and then exploded in the mid-1990s, when the
importance of the Internet became apparent. Politicians could boast that the
network revolution helped revive American fortunes in the computer and com-
puter networking equipment markets.
Shaping the U.S. ICT Architecture 115
8/2/2019 Cowhey Aronson Richards
12/21
But this policy mix had implications for the market structure and infrastructure
architecture that were less appealing. The AT&T divestiture created a logical
segmentation of the physical infrastructure (between long-distance and local trans-
port facilities) that forced the new AT&T to compete against entrants in the market
segments where entry and competitive investments were easiestnot just in
equipment but also in long-distance voice and data services. But it also created amonopoly for much of the national infrastructure and services, local communica-
tions, and, with minor exceptions, froze the regional Bell companies out of the
long-distance and equipment markets. A generation of economists declared,
however, that price performance for long distance would have improved more
quickly if the Bells could compete.47 It also pushed the market in a direction where
there were many new long-distance networks without comparable investment in
local communications infrastructure.
Meanwhile, there was consensus that the Internet and the Web would lead the
next boom in communications and IT investment. The major corporate playerswanted to be ready. The bargain leading to the 1996 Telecommunications Act was
struck between a Republican Congress and a Democratic White House, each with
reasons for wanting to reach an agreement.
Predictably, the politically muscular regional Bells, which operated in all 435
Congressional districts, wanted permission to compete in all markets. Republicans
sided with the Bells because their strength was greatest in the West and the South
where the Bells were especially influential, and the Republicans had won control of
Congress in the 1994 election. By this time, Republicans had won much of the
South, so Democrats (especially the White House) relied on a strong urban base
where local costs of service were lower, demand for consumer long-distance anddata services was high, and large users were concentrated. This tended to align
Democrats with the long-distance carriers.48 The long-distance companies recog-
nized that pressures for Bell entry were enormous, but they counted on Clinton
White House support on the terms for their cross-entry into local services. The
White House did so, but Democrats also were rebranding themselves as the pro-
competition champions of the information economy, so they did not want to oppose
allowing the Bells to compete.49
During the legislative bargaining, the Bells rejected a deal that guaranteed them
entry into the long-distance and data markets three years after passage of the Act.
They thought that they would gain entry faster by meeting a check list of obliga-tions on sharing their network functions with new competitors. However, a Demo-
cratic FCC, with strong White House support, interpreted the Act to call for strong
interconnect obligations for the Bells at long-run incremental costs (a formula
producing steep discounts). This formula enraged the Bells and the Republican
Congress.
Many economists, wary of heavy government regulation, worried that the FCCs
approach might discourage investment by the Bells and induce inefficient, subsi-
dized entry that rested on the Bells unrealistically priced facilities.50 The Bells
launched a full-scale legal counterattack on FCC rules, and as American adminis-
trative bureaucracies enjoy less latitude than their counterparts in parliamentary
democracies, court challenges tied up parts of the interconnection regulation. Still,
market bargains on competitive entry were struck because the Bells wanted to claim
116 Peter F. Cowhey, Jonathan D. Aronson, and John Richards
8/2/2019 Cowhey Aronson Richards
13/21
that they had fulfilled the 1996 Acts checklist. They wanted to be free to participate
in the anticipated boom in fiber optic backbone networks being built in anticipation
of the Internet and e-commerce frenzy. From 1998 to 2002, the demand for
long-distance voice and data quadrupled, an impressive figure, but capacity grew
500-fold.51 This infrastructure later helped to kill the pricing structure for tradi-
tional long-distance carriers voice and data transport offerings.52 The much lowercost of data transport also facilitated further build-up of new web applications. But
this was only part of the policy and architectural story.
The emergence of the Internet, to use a clich, changed everything. The Internet
was transformative because its straightforward route to interconnectivity of all
networks soon overwhelmed the popularity of the existing formats underlying
walled gardens for data networking (such as IBMs System Network Architecture)
and email (like MCI Mail) that could hamper competition.53 By 1994, the Internet
swamped commercial email services, and in August 1995, Netscape went public in
the first initial public offering (IPO) of the boom. In the United States, and to alimited extent elsewhere, new Internet services providers like AOL and MSN and
later large content and e-commerce applications like Yahoo, @Home, and eBay,
sprung up to take advantage of the power and scope of the network.
The ingenious design of the Internet was largely a technology story (although the
control of the design process had an important political element).54 In contrast, the
deployment and use of the Internet, including its rapid emergence as the dominant
way to network data, was closely linked to policy and politics. A critical feature
shaping the growth of the Internet and the Web in the United States, and leading
to decisive American leadership in their rollout, was a pair of telecom policy
decisions with direct roots in earlier policy and political struggles over data net-working. First, the FCC declared that telecom companies had to rent network
transmission capacity to the new email/Internet companies at flat rates and on
technologically neutral terms. This allowed new entrants to substitute Internet
protocols and architecture for approaches favored by traditional telecom carriers.
Second, the FCC declined to allow the phone companies to impose a mandatory
charge for dial-up Internet on a per call basis. Instead, consumers could dial into
a local Internet service as part of their standard phone service. Essentially, this
created flat-rate pricing for unlimited use of the Internet and stimulated user
experimentation with the Internet and the Web. No other country followed suit.
Low costs and technological flexibility were keys to experimentation by both usersand Internet service suppliers.55
The change in consumer networking was part of a broader shift in the ICT
architecture. As already noted, until the 1980s, the computer industry evolved in
response to fundamental technological innovation in the chip industry and as a
result of antitrust policies that unbundled the hardware and software industries.
The introduction of the PC in the early 1980s represented the modular decom-
position of the computer into the major value-added components of chips,
software, and systems design and integration. Here, networking policy became
critical.
A crucial element of the success of the PC, compared with its early fate in Japan,
for example, was the ability to use them in a networked deployment, first over office
LANs and then over national and global networks.56 This dynamic was fuelled in the
Shaping the U.S. ICT Architecture 117
8/2/2019 Cowhey Aronson Richards
14/21
United States by growing competition and innovation in the telecom markets. The
deployment of the PC across the business and consumer landscape eventually
fueled the growth of clientserver architecture in the enterprise, created new
packaged software markets for enterprise (enterprise resource planning, produc-
tivity software) and consumer uses (word processing, graphic design), and defined
the architecture for a whole generation of devices and applications.In the 1990s, the Internet further changed the scope of the network, network
performance, and market-based networked applications, albeit largely still rooted
in the PC architecture.57 In the two years prior to the Netscape IPO in 1995, the
Internet/Webs dominance emerged from rivalry with competing computer net-
working systems. Its triumph came from a technological boost. In 1993, a critical
series of government decisions and the wizardry of the web software began dissemi-
nation to those outside the research community. To begin, the competitive long-
distance networks had ample, cheaply priced capacity in place for the new surge
of data traffic. In addition, since 1984, the FCC had exempted value-added infor-mation services, the precursor to Internet messages, from the access charges that
applied to long-distance traffic. As the Internet began to boom, the Bells quickly
recognized the economic costs of foregoing access charges from Internet traffic.
They pressed to reverse the decision, but the FCC, backed by a White House
devoted to its new technology brand, resisted a formula that the FCC staff estimated
would cost the average Internet user an average of $150 per month.58 The uphold-
ing of the 1984 policy allowed the flat rate pricing that made the Web appealing
to consumers over night.
Technologically, the growth of Internet standards, data protocols, and applica-
tion programming interfaces (APIs) beyond the control of any single platformvendor (such as Microsoft had exercised in regard to Windows) created momentum
for more open APIs.59 A myriad of smaller, more specialized applications also
emerged that built their businesses on powerful, cheaper PCs, broadband network-
ing at the office, and widespread narrowband networking in the home. Changes in
the network architecture and performance were coupled with the continuing trans-
formation of computing by the personal computer and the growth of networked
enterprise computing focused on enterprise resource planning and productivity
suites.60
In short, the Internet cemented the architecture for moving forward in the ICT
infrastructuredecentralized intelligence at the edge of the network (e.g., in ter-minals like PCs) would be critical to planning for networked computing. Once
considered a controversial assumption, after 2000, this premise of the ICT infra-
structure became embedded in policy and technological planning by all major
market centers.61
The networking revolution cemented the dominance of the PC (and later, the
smart terminal) for information technology. But it also created a classic political
economy battle over the control of the value added tied to the PC. If telecom policy
blunted control of telecom carriers over data networking, the politics of information
technology turned over whether the PC could be leveraged to control commercial
advantage on the Web. Thus, in theory, Microsoft might leverage its PC operating
system (a bottleneck facility) to unfairly enhance its competitive Internet position at
the expense of competition and consumer welfare. Worries increased that Microsoft
118 Peter F. Cowhey, Jonathan D. Aronson, and John Richards
8/2/2019 Cowhey Aronson Richards
15/21
would use its Internet browser packaged with Windows to promote its own software
add-ons and content.
The political economy logic of the Microsoft antitrust action tracks the history of
U.S. electronics policy. Many large and small rivals located outside the Pacific
Northwest, especially those in California, began a campaign to capture the attention
of state and federal authorities.62 State actions eventually were picked up by afederal antitrust suit that produced a complex and controversial settlement with
Microsoft.
Economists continue to debate the merits of the suit, but the political economy
of its logic tracked the story of telecom policy related here.63 In addition, the legal
action gave a clear signal that architectures for software and the Web faced major
risks both in the courtroom and from wary customers if they did not extend the
principle of modularity to the software ecosystem. Thus, by the early twenty-first
century, modularity had emerged with a vengeance in the Web 2.0. Even Microsoft
opened its interfaces in ways totally unlike the 1990s.
64
Conclusion: The Path of Technology and Political Economy
Considerable attention has been given to finding which system of government
intervention in technological innovation works best. Here, we concentrated on
a prior analytic step. We asked if political economy played an identifiable role in
shaping one of the largest waves of technological and commercial innovation in
modern history, the revolution in telecommunications and computer networking.
We argued that the amount of competition is only one factorthe particular policy
mix for competition also matters, and this policy mix reflects fundamentals ofpolitical economy.
In the case of the United States and the digital revolution shaping the ICT
infrastructure, we offered three arguments concerning the impact of political
economy and policy on the architecture of the ICT infrastructure. First, it tilted
toward an architectural principle of modularity whose implementing policy norms
influenced the paths of both the telecom equipment, computer equipment and
software, and computer networking markets. Second, it created multiple network
infrastructures for telecommunications when other countries either tried to retain a
monopoly infrastructure or carefully limit the number of competitors. It also created
a pricing system that exempted data flows from the cross-subsidies associated withlong-distance voice traffic. Third, it propelled both a particular architecture for
computing (intelligence at the edge of the network) and the full realization of the
potential benefits of the Internet. The United States did not just embrace compe-
tition in networking traditional telecom and new information services. The policy
mix for competition shaped the architecture of the ICT infrastructure, ultimately
resulting in the triumph of the Internet and the Web in the United States. This
breakthrough reverberated globally and created a truly World Wide Web.
Notes
1 The authors would like to thank the Sloan Foundation Industry Studies Program and the collabo-
ration of The Berkeley Roundtable on the International Economy and The Research Institute of the
Shaping the U.S. ICT Architecture 119
8/2/2019 Cowhey Aronson Richards
16/21
Finnish Economy for enabling two conference panels and a research workshop in which these articles
were developed. The comments from two anonymous reviewers and fellow contributors to this issue
were much appreciated.
2 Zysman and Newman (2006).
3 In this paper, we will treat policy principles as a widely shared view concerning causal dynamics.
We will treat policy norms as the specific premises for particular clusters of policy decisions.
4 More technically, modularity means that components that work together interoperate throughtransparent, nondiscriminatory interfaces. Interoperability requires (1) the technological capability
to build separable inputs at competitive prices; and (2) making design choices that ensure that
interfaces connect seamlessly.
5 See the discussion in the third part of the article.
6 The development of these particular technologies in turn have influenced the trajectory of
growth and utilization of said innovation in other countries, shaped by their own specific political
economic constrains; for two examples, see Dossani and Kenney on India, and Kushida and Zysman
on Japan and South Korea (Dossani & Kenney, 2009; Kushida & Zysman, 2009).
7 We explain how the AT&T monopoly over all terminals attached to the telecom network was ended
later in the text.
8 Smart critiques of claims concerning optimal innovation structures are: Breznitz (2007) and Taylor
(2007, 2009).9 Cowhey and McCubbins (1995).
10 Tsebelis (2002).
11 One congressional check on the power of the FCC is the sharing of some of its powers with other
branches of the government. The most important of these is the shared power over competition
policy with the Antitrust Division of the Justice Department. Given the strength of U.S. antitrust laws,
both political parties are sensitive to the possibility of the rival party politicizing competition policy
when it controls the federal government. As a result, the career officials in the Antitrust Division
enjoy a relatively high level of protection from routine political meddling. Decisions on the general
criteria for when to prosecute are subject to guidance by a political appointee, but the White House
is generally circumspect on antitrust matters.
12 Congress wrote the terms of judicial review in the 1934 Act to shape its implementation (Shipan,
2000).13 The evidence is evaluated in Cox and McCubbins (2007).
14 Hundt (2000).
15 On the impact of institutions on telecom policy, especially federalism and pricing, see Noll and
Rosenbluth (1995).
16 Between 1900 and 1933, as national industrial and network markets took form, state authorities
used antitrust actions to shelter local competitors from national competitors that held advantages
over them. Most senators from these states were wary of nationally dominant firms. Americas
veto-oriented system and Congressional distrust of sweeping regulatory powers dampened
impulses toward national economic planning. Industrial policy that might have concentrated
firms into a few national champions was difficult to pass. It should be noted that fostering
the interests of smaller firms is often confused with a pure economic test of consumer welfare
(Cowhey, 1990a). On the structural differences among states over the sweep of Congressionalhistory, see Brady and McCubbins (2002). On states in early antitrust actions, see Bringhurst
(1979).
17 Electronic switching began to supplant mechanical switches, but the first digital switches did not enter
service until the 1980s.
18 Telecommunications and broadcast required separate transmission networks. Even the introduction
of two new broadcast infrastructures, cable and direct satellite broadcast to the home, were dedicated
specialized infrastructures. The advent of satellite communications services in the 1960s helped long
distance, and, later, broadcast to the home. At first, it did little for data transmission. Fiber optic
transmission began to enter the network in 1977.
19 Senators used hearings to pressure the FCC to make sure that lower interstate rates for national firms
did not harm local in-state rates (Temin & Peters, 1986, pp. 324327).
20 On the evolution of software and computer networks, see Campbell-Kelly (2003).21 Cortada (2005, pp. 33, 8990) reports various estimates of ICT as the costs of the largest banks.
His estimates are in the range of 715 percent of the total costs of the banks. As late as 1992, after
networking costs had declined dramatically from the 1970s, networking costs were 10 percent of the
120 Peter F. Cowhey, Jonathan D. Aronson, and John Richards
8/2/2019 Cowhey Aronson Richards
17/21
total. Interviewing and documents supplied to the authors in the 1980s showed that during the
1970s, networking costs were much higher.
22 Siems and Yucel (2005, pp. 1316). Jacobides, Knudsen, and Augier (2006).
23 Conversation with one of the authors. These changes were part of the broader transition to being a
service economy that eventually made sophisticated manufacturing into a part of service product
schemes, as what happened in ICT (Castells, 1996).
24 Mueller (1997).25 Brock (1994, pp. 6574). Defenders of AT&T decrying the courts are exemplified by Temin (1987).
26 Cowhey (1990a).
27 Large customers sought volume discounts and customized service packages for internal private
networks. Computer services, including networking, were profitable but were on a smaller scale than
today. Sales of IBM computers in 1984 were $22.2 billion. The combined revenue of the top five
computer services firms was $3.4 billion, computed from Datamation figures reported in Chandler
(2001, pp. 118119). The high level of oligopsony in communications use is reported in Cowhey
(1990b).
28 Farrell and Weiser (2003).
29 Eventually, the FCC ordered AT&T to create a separate subsidiary for terminal equipment because
of issues concerning cross-subsidies in the competitive equipment market. The FCC did not think
that these decisions would cause local phone rates to balloon (Brock, 1994, pp. 7998).30 Cohen and Noll (1991). The funding of research leading to the Internet was not an exception.
Nobody saw the Internet as commercial data architecture until late in its deployment.
31 On the U.S. and global policy choices and politics, see Cowhey, Aronson, and Richards (2006).
32 The FCC was not allowing the building of an alternative network so it had to regulate the price
on which competitors could lease capacity from the monopoly network. We thank Gerry Faulhaber
for the point concerning wholesale pricing.
33 The FCC went further than this. During a key phase of the markets growth it restricted AT&T from
offering computer information services. One reason why AT&T eventually agreed to its breakup was
to gain entry into this market.
34 Waverman and Crandall (2000).
35 Brock (1994, p. 118).
36 Coll (1986, pp. 1819, 169171).37 AT&T licensed the transistor to other companies in 1952. In 1959, Texas Instruments and Fairchild
introduced the integrated circuit. In 1971, Intel created the microprocessor. IBM introduced its PC
in 1981. In 1986, Cisco introduced the TCP/IP router. Timeline from Chandler (2001, pp. 262265).
38 In 1963, DEC made its first meaningful impact in the marketplace with its minicomputer that
permitted computers on the factory floor (Chandler, 2001, p. 104). On the politics and economics of
the IBM court battles, see Fisher, McGowan, and Greenwood (1983).
39 By the late 1970s, the Japanese had emerged as the major potential competitor to the American
computer industry. These Japanese firms saw computer networking as a key part of their strategy.
So, it was no surprise that the United States and Japan had bitter trade disputes in the mid-1980s
over the terms for competition in global computer networking precisely because of its importance
for IT leadership (Flamm, 1987, 1988).
40 Breyer (1982).41 See the essays on Congressional action in Shooshan (1984).
42 The classic blow-by-blow account of White House thinking, including why it would not overturn its
own Justice Department, is Steve Colls The Deal of the Century (1986). Colls account shows that the
Justice Department and the federal district court hearing the case were not out of control, as many
critics complained. The critics simply had a losing political argument.
43 All major telecom carriers, including the new entrants, were unionized, so a decline in employment
at AT&T was partly offset by new employees at MCI and other firms. This reduced the resistance of
organized labor, a major constituency of the Democrats.
44 The industry profited from the same antitrust legacy that shaped telecom policy when the Justice
Department in 1953 forced the dominant equipment supplier for cable to divest its network hold-
ings. It had made acquisition of an ownership share into a condition of supply to cable networks.
45 Neucherlein and Weiser (2005), Digital Crossroads, chapters 11 and 12, succinctly analyze broad-cast policy. Fascinating journalistic accounts, with some erratic polemics, are: Keating (1999) and
Robichaux (2002).
46 Noam (2001).
Shaping the U.S. ICT Architecture 121
8/2/2019 Cowhey Aronson Richards
18/21
47 The Bells continued to be cross-subsidized by inflated, but nondiscriminatory, fees charged to
long-distance carriers for local transport. Thus, politicians knew that local prices would be stable,
while long-distance prices continued a faster decline than before the AT&T divestiture (Crandall,
1991).
48 The Democrats also wanted to distinguish new forms of subsidy for consumers from programs
identified with the welfare of the Bells. Despite grumbling from the Republican Congress, the FCC
used its discretion to institute a new fee for telecom services to fund the establishment of Internetaccess for schools, libraries, and hospitals. This was a conscious decision to meet the political demands
to keep service widely distributed to all areas, but the Democrats designed the subsidy so that it went
as much to poor urban neighborhoods as it did to rural areas. Inevitably, coverage of middle-class
areas was part of the political bargain.
49 There were divisions in each party. But the median point of each partys Congressional caucus
was significantly different. Conservative Republicans cast this as enhancing competition by taking
regulatory shackles off the Bells. Clinton Democrats stressed enhancing competition by letting new
entrants attack the local transmission bottleneck on the network.
50 Representative critiques were: MacAvoy (1996), Willett (2002), and Houseman and Sidak (2004).
Neuchterlein and Weiser (2005, chapter 3) offer a nuanced law and economics analysis.
51 There was also investment in competitive local infrastructure, but the boom in long-distance net-
working occurred because everyone overestimated demand and underestimated the speed at whichpricing would decline. It was also clear that entry into long-distance and long-haul data transport had
the lowest regulatory and engineering barriers. Of course, by 2001, the first Internet boom fizzled,
but across America and under the Atlantic and Pacific Oceans, a huge new installed infrastructure of
fiber optic transport remained.
52 The collapse of Global Crossing after it completed massive fiber optic submarine cables linking the
United States and Asia allowed for the outsourcing revolution in India and elsewhere (Friedman,
2005, pp. 103106).
53 The beginnings of internetworking dated from the mid-1980s (Cisco shipped its first router in 1986)
when companies and network providers began to interconnect their networks. U.S. policy changes
in 1991 enabled the commercial use of the Internet, setting the stage for the growth of the Internet
during the 1990s (Bresnahan & Greenstein, 1999, pp. 140).
54 Cowhey and Mueller (forthcoming).55 Bar and others (2000).
56 Cole (1996).
57 In Japan, this innovation storm, driven by lower costs, flexible networking, and user coinvention, was
absent. Japan continued to favor vertical integration anchored on the technological planning of the
dominant NTT. Although Japan also introduced telecom services competition, it limited the impact
of competition by placing all new entrants under a micromanaged price umbrella set by NTT.
Network expansion plans need ministry approval. Because the government wished to sustain its
subsidy scheme for electronics firms, Japan required the licensing of value-added networks. It did
not license a network embracing Internet protocols until 1992 (Greenstein, 2005).
58 This estimate was based on the Bells getting the full equivalent of voice access charges (Hundt, 2000,
pp. 133135, 206207).
59 Amid the notoriety of the Microsoft litigation, an important legal right emerged that allowed softwaredevelopers to reverse-engineer software interfaces to create complementary and substitute software.
This trend, praised by most leading analysts, occurred in both the United States and the EU
(Samuelson & Scotchmer, 2002, pp. 15771633; Greenstein, 1993, pp. 3651).
60 On PCs, see Chandler (2001); Ferguson and Morris (1993). The PCs operating system supported
new packaged, mass consumption, software applications, and spurred enormous investment and
innovation around PC-based software. Declining price/performance rations opened the door to
widespread deployment and the adoption of vast enterprise software packages to manage processes
and data across the enterprise. Packaged software for PCs also opened the way to greater comple-
mentarity of software products, particularly between the Microsoft software platform and specialized
software applications. This created a new set of hardware and software industries focused on the PC
ecosystem (from mice to games to all variety of semiconductors). The emergence of the Internet,
and in particular, a new PC application used to browse content and services, reinforced theclient-server architecture that dominated enterprise architectures (Richards & Bresnahan, 1999,
pp. 336371).
61 See Cowhey, Aronson, and Richards (2006) for evidence.
122 Peter F. Cowhey, Jonathan D. Aronson, and John Richards
8/2/2019 Cowhey Aronson Richards
19/21
62 These issues also were raised over the AOL-Time Warner merger but in that case the operational
remedy was low key, the Federal Trade Commission chose to treat the issue as one of leveraging
bottleneck facilities. See the review by Faulhaber (2003, pp. 7397).
63 Evans, Hagiu, and Schmalensee (2006).
64 Ironically, the changes induced by modularity and competition in the 1990s later changed the
political economy of ICT after 2000. Cowhey and Aronson (2009).
About the Authors
Peter F. Cowhey is the Associate Vice Chancellor, International Affairs at the University ofCalifornia, San Diego. He also holds there the Qualcomm Endowed Chair in Communica-tions and Technology Policy. His research has spanned the politics, economics, and policy ofenergy, telecommunications, trade, and biosecurity. His work has appeared in InternationalOrganization, Foreign Affairs, and International Studies Quarterly. His new book (with Jonathan
Aronson) on the political economy of ICT infrastructure development will be published byMIT Press in 2009.
Jonathan D. Aronson is a Professor at the Annenberg School for Communication and theSchool of International Relations at the University of Southern California. Professor Aron-sons work focuses on international political economy with special attention to trade nego-tiations, trade in services, comparative regulation, international strategic alliances, andespecially international telecommunications.
John Richards is a Research Scholar at UC San Diegos Global Information IndustryCenter. His areas of research focus on the competitiveness and regulation of internationalinformation technology and telecommunications markets. More broadly, he has publishedextensively on the evolution and regulation of international markets in network-basedindustries (telecom, IT, aviation). John held posts in the Stanford Economics Departmentand McKinsey & Company. He currently works at Microsoft in their online services group.
References
Bar, F., Cohen, S., Cowhey, P., de Long, B., Kleeman, M., & Zysman, J. (2000). Access and innovation policy
for the third-generation Internet. Telecommunications Policy, 24, 489518.
Brady, D., & McCubbins, M. (Eds.) (2002). Party, process, and policy. Stanford, CA: Stanford University Press.
Bresnahan, T. F., & Greenstein, S. (1999). Technological competition and the structure of the computer
industry. The Journal of Industrial Economics, 47, 140.
Breyer, S. (1982). Regulation and its reform. Cambridge, MA: Harvard University Press.
Breznitz, D. (2007). Innovation and the state. New Haven, CT: Yale University Press.
Bringhurst, B. (1979). Antitrust and the oil monopolyThe standard oil cases, 18901911. Westport, CT: Greenwood
Press.
Brock, G. (1994). Telecommunication policy for the information age. Cambridge, MA: Harvard University Press.Campbell-Kelly, M. (2003). From airline reservations to Sonic the HedgehogA history of the software industry.
Cambridge, MA: MIT Press.
Castells, M. (1996). The rise of the network society, the information age: Economy, society and culture (Vol. I). Cambridge,
MA; Oxford: Blackwell Publishers.
Chandler, A. D., Jr. (2001). Inventing the electronic century. New York: Free Press.
Cohen, L., & Noll, R. (1991). The technology porkbarrel. Washington, DC: Brookings Institution.
Cole, R. (1996). Telecommunications competition in world markets: Understanding Japans decline. In J.
Zysman & A. Newman (Eds.), How revolutionary was the digital revolution? National responses, market transitions,
and global technology (pp. 125147). Stanford, CA: Stanford University Press.
Coll, S. (1986). The deal of the century: The breakup of AT&T. New York: Athenaeum.
Cortada, J. W. (2005). The digital hand: How computers changed the work of American financial, telecommunications,
media, and entertainment industries (Vol. 2). New York: Oxford University Press.
Cowhey, P. (1990a). States and politics in American foreign economic policy. In J. Odell & T. Willett (Eds.),Blending economic and political theories (pp. 225252). Ann Arbor: University of Michigan Press.
Cowhey, P. F. (1990b). The international telecommunications regime: The political roots of regimes for high
technology. International Organization, 44, 169199.
Shaping the U.S. ICT Architecture 123
8/2/2019 Cowhey Aronson Richards
20/21
Cowhey, P., & Aronson, J. (2009). Transforming global information and communications markets: The political economy
of innovation. Cambridge, MA: MIT Press.
Cowhey, P., Aronson, J., & Richards, J. (2006). The peculiar evolution of 3G wireless networks: Institutional
logic, politics, and property rights. In J. Zysman & A. Newman (Eds.), How revolutionary was the digital
revolution? National responses, market transitions, and global technology (pp. 291323). Stanford, CA: Stanford
University Press.
Cowhey, P., & McCubbins, M. (Eds.). (1995). Structure and policy in Japan and the United States. New York:Cambridge University Press.
Cowhey, P., & Mueller, M. (forthcoming). Delegation, the dimensions of networks and the design of
Internet governance. In M. Kahler (Ed.), Networked global politics. Ithaca, NY: Cornell University
Press.
Cox, G., & McCubbins, M. (2007). Legislative leviathan (2nd Ed.). New York: Cambridge University Press.
Crandall, R. (1991). After the breakup: US telecommunications in a more competitive era . Washington, DC: Brookings
Institution.
Dossani, R., & Kenney, M. (2009). Service provision for the global economy: The evolving Indian e xperience.
Review of Policy Research, 26(1,2), 77104.
Evans, D. S., Hagiu, A., & Schmalensee, R. (2006). Invisible engines: How software platforms drive innovation and
transform industries. Cambridge, MA: MIT Press.
Farrell, J., & Weiser, P. (2003). Modularity, vertical integration, and open access policies: Towards a convergence
of antitrust and regulation in the Internet age. Harvard Journal of Law and Technology, 17, 86134.Faulhaber, G. R. (2003). Policy-induced competition: The telecommunications experiments. Information
Economics and Policy, 15, 7397.
Ferguson, C. H., & Morris, C. R. (1993). Computer wars: The fall of IBM and the future of Western technology. New
York: Times Books.
Fisher, F., McGowan, J., & Greenwood, J. (1983). Folded, spindled, and mutilated: Economic analysis and U.S. v. IBM.
Cambridge, MA: MIT Press.
Flamm, K. (1987). Targeting the computer. Washington, DC: Brookings Institution Press.
Flamm, K. (1988). Creating the computer. Washington, DC: Brookings Institution Press.
Friedman, T. L. (2005). The world is flat: A brief history of the twenty-first century. New York: Farrar, Strauss and
Giroux.
Greenstein, S. (1993). Markets, standardization and the information infrastructure. IEEE Micro, Chips, Systems
and Applications [Special issue on standards], 13(6), 3651.
Greenstein, S. (2005). The evolution of market structure for Internet access in the United States. Manuscript in
preparation, Northwestern University.
Houseman, J. A., & Sidak, J. G. (2004). Does mandatory unbundling achieve its purpose? Empirical evidence from five
countries (MIT Department of Economics Working Paper No. 0440). Cambridge, MA: MIT Department
of Economics.
Hundt, R. (2000). You say you want a revolution. New Haven, CT: Yale University Press.
Jacobides, M. G., Knudsen, T., & Augier, M. (2006). Benefiting from innovation: Value creation, value
appropriation and the role of industry architectures. Research Policy, 35, 12001221.
Keating, S. (1999). Cutthroat: High stakes and killer moves on the electronic frontier. Denver, CO: Johnson Books.
Kushida, K. E., & Zysman, J. (2009). The services transformation and network policy: The new logic of value
creation. Review of Policy Research, 26(1,2), 173194.
MacAvoy, P. (1996). The failure of antitrust and regulation to establish competition in long-distance telephone services .
Washington, DC: AEI Press.Mueller, M. (1997). Universal service: Competition, interconnection and monopoly in the making of the American telephone
system. Cambridge, MA: MIT Press.
Neuchterlein, J., & Weiser, P. (2005). Digital crossroads. Cambridge, MA: MIT Press.
Noam, E. (2001). Interconnecting the network of networks. Cambridge, MA: MIT Press.
Noll, R., & Rosenbluth, F. (1995). Telecommunications policy: Structure, process, and outcomes. In P. Cowhey
& M. McCubbins (Eds.), Structure and process in Japan and the United States (pp. 119176). New York:
Cambridge University Press.
Richards, J., & Bresnahan, T. F. (1999). Local and global competition in information technology. Journal of
the Japanese and International Economies, 13(4), 336371.
Robichaux, M. (2002). Cable cowboyJohn Malone and the rise of the modern cable business . New York: John Wiley.
Samuelson, P., & Scotchmer S. (2002). The law and economics of reverse engineering. Yale Law Review, 11,
15771633.
Shipan, C. R. (2000). Designing judicial review: Interest groups, Congress, and communications policy. Ann Arbor:
University of Michigan Press.
Shooshan, H. M. (Ed.) (1984). Disconnecting Bell. New York: Pergamon Press.
124 Peter F. Cowhey, Jonathan D. Aronson, and John Richards
8/2/2019 Cowhey Aronson Richards
21/21
Siems, T., & Yucel, M. (2005). Where ITs @: Technology and the economy. Federal Reserve Bank of Dallas,
1(January/February), 1316.
Taylor, M. Z. (2007, August). National innovation rates: The evidence for/against domestic institutions. Paper presented
to 2007 APSA Convention, Chicago.
Taylor, M. Z. (2009). International linkages and national innovation rates: An exploratory probe. Review of Policy
Research, 26(1,2), 127149.
Temin, P. (1987). The fall of the Bell system: A study in prices and politics . New York: Cambridge University Press.Temin, P., & Peters, G. (1986). Is history stranger than theory? The origin of telephone separations. In W. N.
Parker (Ed.), Economic history and the modern economist (pp. 324327). Oxford: Basil Blackwell.
Tsebelis, G. (2002). Veto players: How political institutions work. Princeton, NJ: Princeton University Press.
Waverman, L., & Crandall, R. (2000). Who pays for universal service? Washington, DC: Brookings Institution.
Willett, T. (2002). The political economy of cable open access (AEI-Brookings Joint Center Working Paper No.
0009). Washington, DC: AEI-Brookings.
Zysman, J., & Newman, A. (2006). Transforming politics in the digital age. In J. Zysman & A. Newman
(Eds.), How revolutionary was the digital revolution? National responses, market transitions, and global technology
(pp. 391411). Stanford, CA: Stanford University Press.
Shaping the U.S. ICT Architecture 125