United States. Congress. Senate. Committee on the.

The Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) online

. (page 74 of 140)
Online LibraryUnited States. Congress. Senate. Committee on theThe Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) → online text (page 74 of 140)
Font size
QR-code for this ebook


lines. ... As substantive antitrust law dealing with monopolization and multi-
line firms develops, relief which requires separation of defendant firms from
certain markets will continue to be of importance to the reduction of excessive
concentration in the economy." (emphasis added)

In light of the fact that the avoidance of market concentration and dominance
is the central theme of Section 2 of the Sherman Act, the suggested containment
of IBM is both necessary and fully warranted.*

PROPOSALS FOR STRUCTURAL RELIEF

I. Divestiture of IBM's Components Division, Combined with Injunctive
Provisions Against IBM Resuming the Business of Manufacturing Components.

A. The Problem

Components are the elemental parts of a computer, such as wire contacts,
relays, switches discrete semiconductors and, most important, integrated semi-
conductor circuits. IBM purchased most of its components, including semicon-
ductors, until 1958 when it began the manufacture of its semiconductors. IBM
now has the largest semiconductor manufacturing capability in the world, semi-
conductors being the major product of the IBM Components Division. IBM has
seldom if ever made its semiconductors available to other computer manufac-
turers, and IBM purchases only a minimal proportion of its requirements for the
more sophisticated semiconductors from outside vendors. Tlius, IBM semi-
conductors are largely a closed preserve — with IBM neither buying nor selling
the semiconductors critical to the manufacture of computers.

The importance of this kind of vertical integration has been intensified by
a pattern of technological change in which more and more of semiconductors
and their circuitry have been combined on a single chip, so that a single chip now
performs major computer functions. Robert O. Wilson, manager of IBM's Sys-
tems Development Division is quoted as follows : "The keystone to ease of
computer use is large memories and semiconductor technology gives us that."
(Fortune, March, 1972. p. 149)

IBM has been able to manufacture semiconductor circuits at a unit cost that
cannot be matched either by other computer manufacturers or semiconductor
manufacturers. IBM's cost advantage arises from the very large scale of its semi-
conductor operation. Its competitors in the computer industry cannot match
that scale in their own component manufacturing because their individual
market shares in the computer industry are less than a tenth that of IBM.
Control Data and other IBM competitors purchase most of their semiconductors
from such companies as Texas Instruments, Fairchild and Motorola. None of
these suppliers has volumes comparable to IBM's nor the ability to set standards
in the more sophisticated semiconductor devices used almost exclusively in com-
puters. Since IBM's own captive operations account for over 70 percent of the
potential market, refiecting IBM's dominance of the computer industry, the
potential volume of individual suppliers is necessarily limited. Thus, neither
firms in the computer industry nor in the semiconductor industry have the
capability of developing and manufacturing semiconductor circuits at a cost close
to that of IBM. The resulting competitive disadvantage imposed on IBM's
rivals in the computer industry, already critical, is likely to become even more
acute as integrated semiconductor circuits become an even more important
part of the cost of the computer main frame in the coming decade.

It is significant to point out that large scale integrated circuits (LSI) are now
entering the state of development that allows their incorporation in new genera-
tions of computers. Tlie economies of scale associated with mass production of
the LSI circuits will provide both a technological and production economy to
IBM that may foreclose other manufacturers from the computer manufacturing
industry within a decade. Means must be found to allow all computer manu-
facturers to share in these economies of scale so as to provide the consumer
with the benefits of mass production and effective competition.



♦The suggested relief addressed to the computer related markets Is particularly bene-
ficial because such markets are well served by smaller firms which can more effectively
meet the diverse and changing needs of those growing markets than a corporate giant.
This relief furthers the objectives outlined in the President's March 6. 1972 message to the
Congress on "Science and Technology" by improving the climate for innovation and small
high technology businesses.



5529

B. The Suggested Relief

The solution lies in a divestiture of IBM's Components Division which con-
tains its semiconductor development and manufacturing capability. The new
unit would become a separate company able to sell to IBM but required to
make its products available to the other computer manufacturers at prices
corresponding to those charged IBM.

To prevent IBM from recreating this source of monopoly power, IBM should
be prohibited from negotiating with any manufacturer for the purchase of
proprietary circuits. It could, of course, participate in the development of such
circuits with a semiconductor manufacturer, but such circuits should be avail-
able for sale to the other computer and semiconductor manufacturers. It should
also be prohibited from negotiating special prices with any manufacturer for
semiconductors. Finally, it should be barred from resuming semiconductor manu-
facture after the divestiture, either through acquisition or internal expansion
or through importing components from foreign subsidiaries.

C. Evaluation

The divestiture proposed here is clearly feasible. IBM's Components Division
already has a distinct divisional management, separate manufacturing plants,
and its own research facilities.

Since the proposed divestiture does not change the size of IBM's present
Components Division, there would be no loss in economics of scale in these
operations. Nor is it likely that there would be significant loss in "coordination"
economies between IBM's computer operations and its components needs. IBM
could still sponsor the development of components with independent component
companies, including the successor its Components Division. IBM would still
have strong incentives to continue such development, given its massive size and
the critical relationship between progress and economies of scale in components
and computers. What IBM could not do is bar the benefits of such progress
and economies of scale in semiconductors from its competitors in the computer
industry. Finally, there need be no rise in profit margins. IBM now earns a
return on its capital devoted to components — which return would represent the
profits of the new company.

Progress in components would be available to the entire computer industry
and others on an equal basis. One source of IBM's dominance in the general
purpose computer market would be eliminateid and its rivals would be more on
a competitive parity. Additionally, this divestiture would open up IBM's vast
semiconductor requirements to the companies like Texas Instruments, Motorola,
and Fairchild, who could now compete to serve IBM. At the same time, these
companies would have a major new competitor in the successor to IBM's Com-
ponents Division. A vigorous independent U.S. semiconductor industry is, of
course, a boon to the computer industry as well as an essential element in
America's technological leadership at home and abroad.

II. Divorcement of IBM from the Business of Making and Selling Remote
Terminals and Data Communications Equipment.

A. The Problem

Remote terminals and related software have special design features or
architecture that permit use via communications facilities at locations con-
siderably distant from the central processing unit. In addition, there are other
types of equipment — such as modems, controllers and multiplexing units —
that also are communications-oriented in their design. The divorcement pro-
vision would cover all equipments designed and manufactured for intercon-
nection with communications facilities.

These various kinds of data communications equipment have a very high
growth potential in the decade ahead. IBM may soon dominate the remote
terminal and data communications equipment market. IBM already is a major
factor in the mini-computer market, a major component of remote computing
networks. It is also the major seller for many other types of data communica-
tions equipment. Of the main frame manufacturers, IBM alone has a full line
of compatible computers combined with an extensive line of data communica-
tions equipment. IBM can connect mini-computers to its super-computers and
its own data communications equipment to create a massive data processing
network for a large customer including data processing service centers. Most
elements of that system would be its own products. IBM then would be in a
position to engage in activities that have the effect of "full-line forcing", a



5530

classic monopoly tactic, thus locking out IBM's rivals in the general purpose
computer market as well as the independent manufacturers of remote termi-
nals, peripherals, and mini-computers. The support divorcement relief would
accomplish both the prevention of the threatened domination by IBM of this
"terminal" market and the curtailment of IBM's exclusionary power in the
general-purpose computer market.

B. The Suggested Relief

IBM would be divorced from the business of manufacturing and marketing
equipment designed for remote terminal usages, including all data communi-
cation equipment, such as terminals, modems, interfacing controllers and multi-
plexing units.

C. Evaluation

Tlie proposed action presents no feasibility problems. IBM could readily de-
vote its present manufacturing facilities for these products to technologically
similar ones. The remote terminal market can be easily supplied by other manu-
facturers who already have established products and competence in these ac-
tivities. These products also lend themselves to new entry by innovative smaller
companies.

IBM central processing equipment would still be available for operation
with terminals and data communication equipment purchased from other manu-
facturers. Competition among the various terminals manufacturers — both those
now established and among smaller companies recently entering this business —
will encourage the growth of this mode of computer operation which often re-
sults in savings to the users. Indeed, IBM will then have a major interest in
the growth of the terminals business of other manufacturers and will be en-
couraged to work out usable interface standards. At the same time, one prom-
ising method whereby IBM could perpetuate its general purpose computer
system dominance will be blocked.

III. Divestiture of Service Bureau Corporation and Divorcement of IBM from
the Data Services Business.

A. The Problem

The 1956 consent decree implicity acknowledged that one source of IBM's
monopoly power was its practice of starting a customer at an IBM service bureau
for the purpose of familiarizing him with IBM computers and causing the prepa-
ration of his data on IBM media. IBM then used this exposure to, and reliance
upon, its products and personnel to sell IBM equipment. Given IBM's dominant
position, such practices placed its computer-manufacturing competitors at a
substantial disadvantage. The 1956 consent decree required IBM to confine its
service bureau business, i.e., customer data handling and processing (including
facilities management), to SBC, a wholly-owned subsidiary which was to be
operated at arms-length and not to assist IBM's sales activity in computers.

The prescribed "independence" of the wholly-owned subsidiary has not accom-
plished the antitrust objective. SBC remains an aid to IBM's marketing of
computers — in part because SBC uses IBM equipment. As a result, the data of
SBC customers are still prepared on IBM media and to IBM software require-
ments, and so, when those customers subsequently install their own computer,
any choice but an IBM computer would entail a significant cost for the conver.sion
of the data to a new media and re-writing or modifying software. Thus SBC's
activities aid IBM in maintaining its monopolistic position in the manufacturing
of general purpose computers and marketing them to customers including service
bureaus and other data service vendo'-s. (In addition, SBC has been a conduit
for price discrimination by IBM in the sale of computers through free use of
computers, free program support, and buybacks of computer time or programs.)

A broad market for data services has developt^d and includes the following
either separately or in combination :

1. Sales of computer time locally on site.

2. Sale of computer time via comnuinication lines and terminals to a remote
site including what is sometimes referred to as time-sharing (interactive) or
batch (non-interactive) computing.

3. The sale of subscription services involving the use of a computer on a flat
rate or time basis.

4. The sale of facilities management to provide some or all of the data process-
ing services for a particular customer including management, eauipment, pro-
gramming and operating personnel, programs, communications and the space or
buildings for physically housing the operations.



5531

(The sale of data services always involves the use of a computer system and
may or may not include operators, prcgrams, data bases, remote terminal rental
and communications facilities costs.)

The growth of the data services marlcet will intensify IBM's advantages in its
ownership of SBC. Time-sharing or remote computing permits a customer to
achieve the economies and computing power of large scale computers through
using a terminal linked to a centrally located large scale computer used con-
currently by several users. Such a system spreads the costs of a large scale com-
puter over 'many users and for this reason, the demand is expected to grow
rapidly in the coming decades. For example, extensive computer aided instruction
for local schools will become feasible only in a time-shared mode.

IBM, with its dominant position in the manufacture and marketing of general
purpose computers, is in a position to dominate the emerging data services
market. With its full range of products including large central processors, mini-
computers and other remote terminals along with the vast library of network
and applications software, IBM can establish facilities to provide data services
at less cost than competitors. IBM's vast resources would enable them to blanket
the country with networks and centers if they chose.

B. The Suggested Relief

SBC would be divested by IBM and established as a totally independent com-
pany, and IBM barred from re-entering, by acquisition or internal expansion, the
data services business. SBC and other service bureaus may still use IBM general
purpose computers.

C. Evaluation

The feasibility of such divestiture is demonstrated by the fact that IBM has
been required to operate SBC as an independent "arms-length" subsidiary since
1956.

A divorcement of IBM from the data services market including a totally inde-
pendent SBC is the only way to insure that marketing data services does not
assist IBM in maintaining its monopoly position in the sale of general purpose
computers. An independent SBC, combined with a prohibition against IBM's
directly engaging in providing computer services, would also remove the major
threatof IBM's domination of the data services market. In addition, SBC might
become a substantial equipment customer for IBM's rivals in the general purpose
computer market.

IV. Divorcement of IBM from Professional Services and Education Activities,
and Divestiture of SRA.

A. The Problem

The manufacture and marketing of hardware (main frame computers and peri-
pheral equipment) will continue to be the most important segment of the indus-
try for the foreseeable future — for they are the heart of any computing system.
However, hardware now represent less than 507o of the users expenditure for
data processing. Computer training or education and professional services, will
become even more important in the coming decades.

By professional services we mean consulting and programming services on a
contract or fee basis. Consulting services includes systems analysis and design,
systems engineering and specifications, feasibility studies, systems selection,
architectural design, and similar activities. Programming services include the
designing and implementing of programs for the application of computers includ-
ing application design and development, program design and development, pro-
gram conversion, testing and implementation as related to specific user applica-
tions. Program products developed internally at IBM and licensed for a fee —
including operating systems, computers, applications programs, and utility
routines — are not included in this definition.

Education includes training of operators, programmers, maintenance and
management personnel in the computer field.

The user's experience with a particular manufacturer's computer is closely
related to the marketing of general purpose computers. IBM has capitalized on
this linkage by offering massive educational discounts on computer systems to
major universities. The importance of the major universities in influencing
the computer purchase decisions of large groups of users will likely be matched
in the future by other kinds of computer education and training. While the train-
ing of key scientists remains important, industry trends indicate the growth of
educational courses now offered by commercial schools in major urban areas to



5532

train the large number of highly skilled technicians and to provide refresher
training for these technicians to work effectively with the rapidly changing tech-
nology. IBM already has a major edge in such training because most commercial
schools use IBM equipment. IBM also has begun to offer .such training directly,
including the operation of its System Science Institute.

Any expansion into the education market by IBM with its massive resources,
could well be a key element in perpetuating IBM's market dominance in the
general purpose computer market. Thus, IBM's participation in the educational
market in the '70s and '80s could substitute for its massive educational dis-
counts of the '60s. The relief proposed is addressed to preventing this way of
perpetuating dominance in the general purpose computer market, as well as
preventing the potential dominance of the educational services market by IBM.

Professional services present an analogous problem. Until recently IBM often
gave away such services and realized a return on these "free goods" through
continued dominance of the general purpose computer market. Now that many
users have well established computer installations, services that add to the
efficiency and extend the range of computer usage will be more important. Com-
puters are being applied to problems requiring a higher order of expertise which
call for more reliance on highly-specialized consultants and programmers.

Now that unbundling creates the prospect of a free market in professional
services, and this market is likely to grow rapidly. IBM may attempt to domi-
nate this market to further enhance its sales of general purpose computers. Its
resources, including the fact that it has the largest number of computer profes-
sionals of any company in the world, provides the potential for IBM to dominate
the professional services market. Again, the relief proposal serves to loosen
IBM's hold on the general purpose computer market, as well as to prevent IBM
dominance of this rapidly growing service market.

B. The Suggested Relief

IBM should he divorced from the business of providing educational and pro-
fessional services. IBM would be allowed to continue to engage in the training
of its own employees, and to offer training directly associated with the sale or
lease of IBM computer systems, but these latter services should be clearly speci-
fied at time of sale and limited to those associated \^dth the initial operation of
newly acquired equipment.

Inasmuch as the business of SRA is closely related to the educational services
market, that wholly-owned subsidiary, which has been operated independently
by IBM since its acquisition, should be divested.

C. Evaluation

These service.s lend themselves fo provision by a large number of competitors
which (subject to potential exclusion by an exercise of IBM's powder) are al-
ready fully capable of providing the required services. In particular, the inde-
pendent software houses would benefit if IB]M were prohibited from using its
hardware dominance as a wedge to usurp the position of such companies in de-
veloping software. The continued viability of these independent companies, and
the continuation of competition in these markets, can be insured by divorcing
IBM from these businesses and confining IBM to the principal business of
manufacturing and marketing computer systems. No feasibility problem presents
itself with respect to such relief.

V. Divestiture of IBM's Office Products Division and Divorcement of IBM
from the Type of Business Conducted by that Division.

A. The Problem

This IBM division develops, manufactures, markets and services computer-
related (or potentially computer-related) office products, the most important of
which is typewriters. IBM is the dominant manufacturer of office electric type-
writers, having 86 percent of the heavy-duty office typewriter market according
to the February, 1972 initial decision of the FTC hearing examiner in the Litton
Industries case. Given the universality of the use of the typewriter, this product
gives IBM salesmen and maintenance men. as well as its office systems and pro-
cedure personnel, access to almost every office in the country.

Typewriters are used as data preparation devices for entry of data into com-
puter systems. This can be accomplished via optical character recognition type
fonts on the typewriter or through auxiliary equipment that uses the typewriter
to produce machine readable media, such as paper or magnetic tape. Typewriters
are also used as output devices to print hard copy after computer processing.



5533

In addition, typewriters are used for computer consoles. Further, as higli-
liglited in IBM's 1971 annual report, IBM typewriters are a major element of
the i-emote terminals made by IBM and those made by its competitors as well.
Almost all the computer manufacturers, therefore, are dependent to some extent
on IBM typewriters. Moreover, office copiers, a product line IBM recently en-
tered, are. with some modification, beginning to be used as computer output de-
vices and could serve ;is a further menus oi' tying the customer.

IBM's commanding position in typewriters helps to maintain its dominant
position in the general purposes computer market in several ways : (1) IBM has
an entry into almost every small business in the country, most of which are poten-
tial, if not present, computer users; (2) IBM can learn customer business sys-
tems and procedures, and educate the customer to IBM's "way of thinking"
so as to condition them to the use of IBM computers; and (3) typewriters and
office copiers provide a springboard for the design and sale of comprehensive in-
formation systems that are likely to have a major computer application in the
future.

Conversely, the sale of computer systems often requires or includes a large
numl^er of typewriters for data entry, either directly connected as terminals or
to prepare data for optical character reader entry.

B. The Suggested Relief

IBM's Office Products Division would be divested and established as an inde-
peiK'ent company. IP>?.[ would be b;irred from re-entry into this line of business,
through acquisition or internal expansion.

C. Evaluation

Tlie Office Products Division is operated as a separate entity, with its own
research, manufacturing plants, management, and sales operation. There is no
question as to the economic feasibility of divesture. Office Products would be a
successful independent company — in fact, the leading company in typewriters.

l-'reed of IBM ties, office products would no longer be a marketing asset which
IBM can u.se in the computer markets in an anti-competitive way. These prod-
ucts would be freely available to other manufacturers and at non-discriminatory
prices. As an additional benefit, IBM's monopoly or near-monopoly position in of-



Online LibraryUnited States. Congress. Senate. Committee on theThe Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) → online text (page 74 of 140)