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 62 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 62 of 140)
Font size
QR-code for this ebook


— before the solution which the Justice Department woukl like to im-
pose would be developed, screened, analyzed, adjudicated, negotiated,^
and implemented.

At the very best, this means that a solution is created which solves
the problem identified 10 to 20 years previously.

In a high-technology, rapidly changing industry this is obviously
unacceptable. Justice dehiyed is justice denied; and justice delayed is,
in this particular case, if it is to happen, to be applied to a situation
fi'om ancient history.

Given the legal resources of a large monopoly, the necessity to
operate within our current laws, and with the safeguards of our cur-
rent legal system, it appears that reasonable speed in the solution
of a monopoly problem is not possible today.

The problem of a timely solution can be divided into two parts:
Tlie first is. is it possible within our CTirrent laws to establish a viola-
tion of the Sherman and/or Clavton Act on the part of a dominant
companv in a reasonable period of time ?

It's clear that one purpose of S. 1167 is to solve that problem
by establishing criteria more easily proven.

T commend that concept, if not the specific criteria of that bill.
The second problem is, given the finding of monopolization, is it pos-
sible within our current structure to establish and carry out a relief
plan that will solve the problem ?

From past history one has to assmne that the courts, upon prosecu-
tion by either the Federal Trade Commission or the Justice Depart-
ment, have an automatic l)ias toward regulation.

It seems quite straightforward and easy for a court to lay down a
series of injunctions, either by court order or through a consent
decree, which regulate specific practices or business conduct.



5453

From the few instances in which a sohition has been found through
restructuring- or divestiture we must assume tliat our current sj'stem
is nuich less able to deal with any solution other than regulation.

I submit that it's patently clear that current laws do not provide a
timely solution to establishing violations and correcting structural
problems.

In the particular case of the computer industry, what criteria should
a solution meet '^

The first and most important is that the solution should break the
market power of the dominant company or control it in such a way that
it is deleterious neither to computer users nor other competitors at-
tempting to supply alternate products.

This inchides insurance that IBM or its remaining entities do not
have the unilateral ability to create and change the enviromnent in
the marketplace.

To accomplish this it would undoubtedly be necessary to eliminate
or sterilize the $4 billion in cash assets that stand behind IBM's
enormous financial power in market control.

You notice my number is a little larger than some of the other peo-
ples because theirs was as of the end of last year, and I have adjusted
it for the cash infusion since then.

Any solution would have to lead toward an interchangeability of
competing systems. This implies a rigorous industry effort on stand-
ards and interchangeability in both software and hardware.

Only if such compatibility were established would it be possible for
competing entities to offer comparable products and services to all
computer users.

Assuming that a solution did not create a large number of competing
entities in place of IBM, it would be necessary that no company coulcl
dominate the injection of technology bv surprise announcements of
new interfaces between various pieces of hardware and between hard-
ware and software.

Next, any solution would have to encourage, or at least not badly
discourage, entry of new firms. If we define a competitor as a firm that
is 1 percent of tlie size of IBM then tlieie liave l)een no new competi-
tors in the computer industry for at least 10 years.

If we define success, as many industries clo, as 10 percent of the
market, there lias never been a successful competitor.

The computer industry today contributes a significant dollar amount
to our balance of trade. A solution must not change, and in fact should
be designed to improve, the positive balance of trade created b}' the
com.puter industry.

Perhaps most important!}^, any solution should take advantage of
the forces of free enterprise and managerial entrepreneurship which
are available to tliisecoiioniv.

And last, a solution should be implemented this century if not this
decade.

The boundaries within which a solution will be found are the
restructuring of IBM into several competitive entities, as requested by
the Justice Department; the regulation of IB]\I"s business and business
practices in n form comparable to the consent decree of IDoB: f^nd a
whittling of IBM's market share through a combination of regidation
and divestiture, looking forward to a time when more competitive

40-927 — 75 40



5454

conditions would exist in the industry — at which time the regulations
could be removed.

It is my miderstanding that witnesses tomorrow will cover some,
if not all, of these possible solutions. But I would like to comment on
them in advance.

Regulations through court order, whether arrived at by a judgment
or a consent decree, would lay down a set of rules under which IB]M
would have to conduct its business in the future.

The rules devised would be intended to guide and control IBM's
market practices and policies for the benefit of users, and to permit
a viable competitive industry to develop and prosper.

The rules would undoubtedly include some restraint on the per-
mitted contractual terms for IBM to do business, perhaps, for example,
prohibitino- leases of greater than a dui'ation of 1 year.

It would also include rules under which pricing could be set and tech-
nological changes introduced. The Computer Industry Association
submitted a recommendation to the Justice Department 18 months ago
for iust such a set of rules.

However, our goal at that time was that those rules would be to
regulate the industry only imtil the Justice Department case was
completely settled.

There is ample precedent for this kind of action in the U.S. econ-
omy. "Wliere problems of natural monopoly or unlimited economies
of scale have arisen in the past, this solution has been often used.

On the other hand most of the current trends within both the execu-
tive and legislative branches of the Government are directed away
from such regulations.

In fact there are numerous efforts afoot today to deregulate some
of the regulated industries.

Most schools of thought, liberal or conservative, tend to believe that
the advantages of the American free enterprise system are reduced
insofar as such Government regulation is undertaken.

The second node of the three nodes of possible solutions, I call
"whittling." It's a combination of temporary regulatory provisions
combined with strictures that require reduction in market share to
some predetermined level or divorcement from activities — such as
terminals or software — which may represent the largest future
growth areas in the industry.

The combination of these two look toward the time when the mar-
ket share goals have been reached and the regulations are then re-
moved.

Since this solution has the expectation of removing all regulations,
either after a given period of time or when a specific event occurs,
they are, to conventional economists, more attractive than indefinite
regulation.

In addition, since it insures an expanding market share for com-
peting companies, this plan has considerable attraction to those com-
panies currently attempting to capture an increased share through
the working of a free enterprise system.

Most of the largest companies in this industry believe their inter-
ests would best be served by such a solution.

On the other hand, the consent decrees won by the United States
against IBM in 1934 and 1956 were of such a kind.



5455

Since the}' liave been largely ineffective in achieving their objec-
tives there are many within the industry and within the Government
Avlio are not sanguine concerning the likely results of such an action.

In an industry as dynamic as data processing it would require
an inordinate degree of foresight to be confident that the expected
results could be achieved.

1 think Mr. Sanders" presentation just prior to mine, where he
pointed out that if j'ou solve the j)roblem in these terms today, the
jump into a total network of communications and computing — where
what we have been talking about here as the systems lock — becomes
a total information lock — would tend to indicate that if we solve to-
day's problems without considering that, it wouldn't solve the prob-
lem in the longer run.

This whittling solution is clearly the most penalizmg one to the
corporate entity of IBM. If it is successful in achieving its objectives
the plan undeniably rewards success by insuring that growth will be
zero, or negative, in the future.

The restructuring solution currently advocated by the United States,
according to their tilings with the court, is intended to go to the heart
of the market power of IBM, splitting it into a number of competing
entities.

Since none of these entities would have the image, installed base,
resources, or monopoly position of IBM, the assumption is that they
would vigorously compete with each other as well as with other firms
in the industry and thereby restore open competition without regula-
tion.

This restructuring would make the objectionable market practices
less likely, since their effectiveness depends upon the total market
power of IBM.

Without such power the encouragement of renting, lack of informa-
tion on interfaces, tie selling, refusal to deal, and unilateral establish-
ment of standards, would be more difficult, if not impossible, to carry
out.

The disadvantage of this solution is that the increased competitive
atmosphere within the marketplace, and the probable lowered cost to
the user, could put the survival and viability of other main-frame
finns in serious question.

By leaving the installed user base intact, even though split among
several firms and dedicated to what is now IBM equipment, additional
market share for CDC, NCR, Honeywell, Univac, and Burroughs
would be almost as difficult to achieve as it is today.

Other factors in the industry would, of course, have additional po-
tential OEM opportunities and OEM competition from the new en-
tities.

Most economists, I believe, would tend to support the Government
solution as the one most likely to bring the advantages of the free
enterprise system to the computer market.

A combination of enlightened self-interest and doubts concerning
some of the projected results of such a restructuring causes many in-
dustry executives to prefer either of the other two nodes of the solu-
tion matrix to the restructuring one.

On the more general topic of your legislation. Senator, I'd like to
advance a few modest proposals which might be worth considering
as additions to S. 1167, or as part of a new piece of legislation.



5456

My comments are focused toward attem])tJiio; to have a solution to
the concentration problem assisted rather than opposed by the con-
siderable entrepreneurial resources available in our f ree enteq^rise
economy.

In the first place, any statute on the subject should establish very
clear criteria of concentration and/or monopolization.

Today it's almost impossible to get a clear statement of what con-
stitutes a violation of section II of the Sherman Act.

Such a criteria could easily be based on market share. AVithout a
market share of appi'oximately 30 percent, held for a reasonable
period of time in a relevant marketplace, there cannot be market
power.

At the same time a marlcet could have a ''four-firm concentration
ratio" of 100 percent with no firm having as much as 30 percent market
share, which would indicate that the criteria is not overly restrictive.

If it can be proven that over 30 percent of a market is required to
achieve economies of scale, there must be firm provisions that protect
the consumer in such an unusual case.

The point is that the danger of monopoly is the holding of market
power rather than something vague like "'attempts to monopolize.*'

But wouldn't such a stricture create a strong disincentive to man-
agement? Why strive and compete if there is a mandated ceiling
beyond which an enterprise cannot grow in the marketplace ?

This is a serious and real problem. ^Monopoly or a significant power
over the marketplace is the goal of all managers of free enterprises.

However, when such a goal is achieved the entrepreneurial energy is
diverted to maintaining and extending the monopoly and avoiding
prosecution by the authorities rather tlian creating ways of increasing
service to customers.

As a consequence an alternate course must be devised which permits
the entrepreneur to continue to compete, and creates greater rather
than less competition.

I'd like to make an additional proposal in this regard, which should
be analyzed and investigated as one possible idea on the problem.

The concept is that of a restructuring tax credit comparable to the
investment tax credit. A publicly helcl company which divests of a
subsidiary completely — setting up an independent company, trans-
ferring appropriate assets to that company, establishing an inde-
pendent board of directors and distri])uting the shares to its stock-
holders — would be permitted to take a credit against income tax for the
following 10 years.

The credit would be a percentage of the assets transferred to the
neAv independent company.

In other words, a successful entrepreneur who reached the ceiling of
his allowable market shares could spin off a competitive entitj- and
achieve not only new ownership in a new enterprise for his stock-
holders, but in addition receive credit in his future income taxes for
the economically sound act which sucli divestiture would represent.

Obviously, economic studies have to be made of the cost of such a
proposal as well as the proper level of tax credit. I would suggest that
tlie cost of such a program could be covered by a graduated income tax
for corporations which increased to something like 65 percent rather
than 48 percent for all taxable income over $100 million.



5457

Obviously the exact percentages in dollars for these proposals will
need to be closely scrutinized, but the concept is to tax to a higher
degree those large agglomerations of power, about which social policy
is properly concerned, in order to pay for the deconcentration insti-
tuted as an entrepreneurial venture by business managers.

There is an almost unlimited store of creative enterprise available
within our economy. Let us structure an environment which harnesses
that energy to achieve deconcentration by incentives rather than
through the dead hand of regulation.

In summary, Senator, w^e have a real problem in the computer in-
dustry. Our current laws appear incapable of solving the problem in
terms of either defining it or creating a constructive solution.

What we need is clearly a criteria for a solution and more creative
ways in which such solutions can be structured.

Thank you.

Senator Hx\rt. Thank you very much. INIr. O'Leary ?

Mr. O'Leary. Let's assume reorganization has been made to make the
industry more competitive. I'd like to get the reaction of someone with
experience in the industry with respect to the feasibility of some of the
ideas which have been propomided in terms of relief.

First, the idea which you touch on briefly in your statement about
making four, or wliatever the number, completely integrated computer
system manufacturers. I am ignoring for the moment some of the
related products which IB^I might manufacture, on components, CPU
systems, software, applications of software, and peripheral equip-
ments. First of all, is it feasible?

Could we create a number of different integrated manufacturers out
of the IBM facilities ; and if it is feasible then would it be wise public
policy ?

Mr. McGiTtK. Mr. O'Leary. I think it's feasible. It's not a simple
job. For example, as the witnesses yesterday pointed out, IB^M has in
Europe carefully arranged so that in each company there is a nonin-
terrupted manufacturing plant; and I believe some of the witnesses
indicated that they believe that that was a necessary economy of scale.

From my experience in the computer industry that is not a neces-
sary economy of scale. The assembly and manufacture of computer
main frames or peripherals is not a heavy capital investment industry.

The manufacturing equipment is not a heavy capital expenditure
relative to the volume of manufactured product. The largest expendi-
ture is test equipment, so I see no problem in organizing some number
of companies around the plants that are currently in existence to have
a number of integi-ated self-sustaining companies.

The exception would be the components division that was discussed
by ^Irs. Carlson. And as she pointed out : If that were spun off, that
is a very vial)le, sound, strong entity.

I think it would be necessary that that be spun off in order for these
other entities to be wholly integrated and self-sustaining.

I might say they wouldn't be completeh' balanced. They wouldn't be
identical. Some might have a stronger capability for disk manufac-
turing, for example, where the others might have only an embryonic
one, or a less capable one, or only a particular kind of disk.

Mr. O'Leary. Some might make big machines; some might spe-
cialize.



5458

Mr. McGuRK. That would be one way to most easily do it, Tliey
could be fully line intergrated. You immediately ask the question,
would they be viable competing across the line i You heard earlier
that there are seven companies considerably smaller than any of those
that would compete across the line.

Mr. O'Leary. So you would have a company which manufactured
components. I would assume that they would have to supply in nondis-
criminatory terms components to these other computer systems manu-
facturers and 3-0U would have four or five or whatever the feasible
number is.

Now, what happens to the remaining sj^stem manufacturers that
are presently in the industry? Do they get cheated by the new firms
which come out of IBM ?

Mr. McGuEK. One can't predict, of course, what might be the course
of events. They would not be under the huge overhanging image of
IB]\I, which is a very important factor in the marketplace.

]Many of the system locks, the uses of the large market power that
you've heard about here, would not be availal^le to those individual
companies.

I would expect none of them would be called IBM so I think that
the better of those new ex-IBM entities and the better of the current
systems manufacturers would sui-^dve and prosper.

I can't say which ones they would be. That would depend largely
on management and, as in all enterprises, a certain degree of luck;
but I see no reason why they couldn't be viable competitive entities.

Mr. O'Leary. ]Mr. Collins talked before about the lower barriers
to entry to the plng-compatible manufacturers who didn't have to
offer the entire system to somebody.

Xow, if you have four or five entities which came out of IBM and
the remaining system manufacturers, have you really lowered the
barriers for those people who want to specialize with respect to termi-
nals, tape drive or disk drive, or do they still have the same problem ?
Is Telex any better off if you have 10 system manufacturers out there ?

]Mr. McGuRK. "Well, there's tAvo sides to that, of course. One of the
big advantages, when a plug-compatible manufacturer designs a prod-
uct, the obvious first interface that he wants to have for that product
is with IBM, because they are 70 percent of the market. So if you
are successful in getting a compatible product you immediately have a
very large marketplace.

The disadvantage is that as they have learned that compatibility
can disappear overnight. In the situation were there are now four
or five entities, there is no one 70-percent plug that you can design
to plug into, but you do have some expectation that none of the others
have the market power to completely unplug you ; and if one does the-
other won't. But more importantly, were that to happen there would
be a real need, and I really believe an important driving force, to-
ward real standardization in the industry.

Xow, standardization docs not meaii making all things alike. It
means making the ability to connect together things and to interrelate
data from one installation to another, if possible.

Standardization is like trailer hitches. It's not like designing cars
alike. It's like having a standard interface so that if you build a
trailer you can hook it on to one of several cars because there's a stand-
ard size ball hitch that vou can attach to.



5459

Xow, it's to the advantage of both the trailer manufacturer and
the car manufacturer to get that standardized.

More importantly, in "our industry customers want the ability to ex-
change their machine data — tapes, disks, and so forth — between in-
stallations. The desire to have compatible media so that you can read
a tape on the Univac machine that was written on an IBM machine is
one of the few areas of real standardization we have today because the
usei"S demand it.

The standard is whatever IBM says, but at least it is a standard
that by now is reasonably accepted. New media come along but the old
]nedia are still standard. I believe if the market was as fragmented as
we're describing, with nobody with, say over 15 percent, there would
be considerable pressure from the industry members themselves — •
and certainly from the users— to get a higher degree of interface
standardization which would then make the esLsy entry markets like
plug-compatible peripherals and terminals much more easil}^ assured.

Mr. Sanders described his need to have an interface so that his new
system will match IBM's. In an industry that is not so dominated,
that would have been worked out by an industry standards group
some time before.

Mr. O'Leary. The other way which is commonly discussed would
be to take different functions of the fully integrated computer sys-
tem manufacturer and see which of those can be pulled apart without
disturbing the significant economies of scale that exist.

Now, what about the CPU and the systems software ? "Would those
two things have to be under the same roof? I think the gentleman
from Auerbach who testified the first day left the impression that
sometimes you create the systems software and then build the hard-
ware around that, which would lead to a possible inference that it
should be the same.

]\rr. ]McGuRK. Let me just comment on that. The only com]~)any that
I Ivuow that has done that is Burroughs. I think that IBM is doing
that for their new system. That is, design the software first and then
desigji the hardware to match it.

It is technically the proper way to go and it is, therefore, I think,
important that a systems design organization — and we're assuming
liere that we have broken that off, the systems design and manufac-
turing the main frame as being a separate entit}' — have the ability and
understandiug to create the systems software.

That doesn't mean to say that it has to be indelibly bundled with
the main frame. One of the — you might say— valid complaints of
IBM in the plug-compatible arguments is that we create all of this
operating systems software, we have to advertise it over everything,
then these parasites come along and plug in and they don't have to
develop that systems software so obviously they could underprice
us.

That's true; but you ought to talk to a software house that tries to
underprice IBM on their S3^stems software. You can't get below zero
and there's no reason that IBM could not charge for their systems
software.

In fact, in ]\fr. Biddle's submission, one of his appendixes was a
recent editorial by a leading executive and designer in a software
house who said if IBM would charge for their systems software we
would see all kinds of innovations.



5460

There would be lots of competitive products in the systems soft-
ware area. He says that. I'm not a software expert. It would seem
difHcult to me, but I don't see any reason why it shouldn't be left as a
possibility.

Mr. O'Leary. What would your reaction be to a proposed reorgani-
zation which would say IBM has to spin off and make a separate cor-



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 62 of 140)