[Senate Hearing 106-596]
[From the U.S. Government Publishing Office]
S. Hrg. 106-596
EXPORT CONTROL IMPLEMENTATION ISSUES WITH RESPECT TO HIGH-PERFORMANCE
COMPUTERS
=======================================================================
HEARING
BEFORE THE
COMMITTEE ON
GOVERNMENTAL AFFAIRS
UNITED STATES SENATE
ONE HUNDRED SIXTH CONGRESS
SECOND SESSION
__________
MAY 26, 2000
__________
Printed for the use of the Committee on Governmental Affairs
_______________________________________________________________________
For sale by the Superintendent of Documents, Congressional Sales Office
U.S. Government Printing Office, Washington, DC 20402
U.S. GOVERNMENT PRINTING OFFICE
65-171 cc WASHINGTON : 2000
COMMITTEE ON GOVERNMENTAL AFFAIRS
FRED THOMPSON, Tennessee, Chairman
WILLIAM V. ROTH, Jr., Delaware JOSEPH I. LIEBERMAN, Connecticut
TED STEVENS, Alaska CARL LEVIN, Michigan
SUSAN M. COLLINS, Maine DANIEL K. AKAKA, Hawaii
GEORGE V. VOINOVICH, Ohio RICHARD J. DURBIN, Illinois
PETE V. DOMENICI, New Mexico ROBERT G. TORRICELLI, New Jersey
THAD COCHRAN, Mississippi MAX CLELAND, Georgia
ARLEN SPECTER, Pennsylvania JOHN EDWARDS, North Carolina
JUDD GREGG, New Hampshire
Hannah S. Sistare, Staff Director and Counsel
Christopher A. Ford, Chief Investigative Counsel
Mark T. Esper, Professional Staff Member
Joyce A. Rechtschaffen, Minority Staff Director and Counsel
Laurie Rubenstein, Minority Chief Counsel
Darla D. Cassell, Administrative Clerk
C O N T E N T S
------
Opening statements:
Page
Senator Thompson............................................. 1
Senator Lieberman............................................ 3
Prepared statement:
Senator Akaka................................................ 39
Witnesses
Friday, May 26, 2000
Harold J. Johnson, Associate Director, International Relations
and Trade Issues, National Security and International Affairs
Division, U.S. General Accounting Office....................... 4
Robert J. Lieberman, Assistant Inspector General for Auditing,
Office of the Inspector General, U.S. Department of Defense.... 6
Gary Milhollin, Director, Wisconsin Project on Nuclear Arms
Control........................................................ 8
Dan Hoydysh, Co-Chair, Computer Coalition for Responsible Exports 11
Alphabetical List of Witnesses
Hoydysh, Dan:
Testimony.................................................... 11
Prepared statement........................................... 102
Johnson, Harold J.:
Testimony.................................................... 4
Prepared statement with an attachment........................ 40
Lieberman, Robert J.:
Testimony.................................................... 6
Prepared statement........................................... 57
Milhollin, Gary:
Testimony.................................................... 8
Prepared statement with attachments.......................... 76
Appendix
Insert for the Record: Response from Mr. Johnson to Senator
Lieberman's Question........................................... 119
Responses to Questions for the Record submitted by Senator Akaka:
From Mr. Johnson............................................. 121
From Mr. Milhollin........................................... 122
From Mr. Hoydysh............................................. 123
EXPORT CONTROL IMPLEMENTATION ISSUES WITH RESPECT TO HIGH-PERFORMANCE
COMPUTERS
----------
FRIDAY, MAY 26, 2000
U.S. Senate,
Committee on Governmental Affairs,
Washington, DC.
The Committee met, pursuant the notice, at 10:09 a.m., in
room SD-342, Dirksen Senate Office Building, Hon. Fred
Thompson, Chairman of the Committee, presiding.
Present: Senators Thompson and Lieberman.
OPENING STATEMENT OF CHAIRMAN THOMPSON
Chairman Thompson. The Committee will come to order,
please. I welcome everybody to this hearing of the Committee on
Governmental Affairs.
Today we are holding a hearing on export control
implementation issues with respect to high-performance
computers. High-performance computers represent a special
challenge for our export control regime, because in many ways
they are the king of dual-use technologies; that is,
technologies that are subject to national security export
controls because they are easily usable for important civilian
purposes as well as dangerous military purposes.
High speed computing, of course, is vital to today's
knowledge-based economy. Unfortunately, however, as the Cox
Report reminded us, powerful computers are also vital to such
things as nuclear weapons development, the design and testing
of ballistic missiles and advanced conventional weapons,
intelligence analysis, code-breaking, military command and
control, and cutting-edge warfare applications, such as
computer network attack.
This is why high performance computer export control issues
are so important. We have to find an appropriate balance
between promoting commerce and protecting our national security
through export controls. If we get it wrong, however, we either
strangle a crucial sector of our information-age economy or we
help potential adversaries prepare to defeat our military
forces in the field, hold our cities hostage to weapons of mass
destruction, or cripple our government and economy through
information warfare.
The debate over high-performance computer export controls
is particularly important in the Senate this year because of
two pieces of pending legislation that affect this balance
between commerce and security.
First is the Banking Committee's proposed reauthorization
of the Export Administration Act, which appeared briefly on the
Senate floor in March. Of most direct relevance to computer
export controls, this bill would have written categories of
``foreign availability'' and ``mass-market'' status into the
U.S. export control law.
That law would require that any control items meeting these
definitions--mass-marketing, foreign availability--be made
available for export without a license to essentially anyone in
the world.
The second pending piece of legislation is a proposal to
shorten the current 180-day period which Congress has in order
to review administration decisions to decontrol computers at
certain performance levels--which are usually measured in terms
of MTOPS, or millions of theoretical operations per second.
Both pieces of legislation are supported by U.S. computer
exporters, but both have also raised serious concerns in the
minds of officials concerned with ensuring that our national
security export controls really do protect national security.
Our discussions today about high-performance computer export
controls will help inform the Senate's consideration of this
and other legislation.
So I hope our discussions will help illuminate a number of
subjects today, but there are a few that I think are
particularly important. First, is it possible, clearly and
objectively, to make the kind of foreign availability and mass-
market status determinations that the computer industry wants
to make the basis of removing controls on many high-performance
computers?
Second, according to what criteria have decisions to
decontrol high-performance computers been made in the past, and
how sound has their analysis been?
Third, even if coherent and objective foreign availability
and mass-market status determinations are possible, who should
make them? Should this be left to the unilateral discretion of
the Department of Commerce, or should our national security
community, such as the Defense Department, have to agree to
decisions to remove export controls of high-performance
computers?
Fourth, if foreign availability and mass-market status
decisions are inherently subjective, and especially if they are
left solely in the hands of the Commerce Department, is it wise
to reduce the congressional review period for such
determinations? At what point would a shortened review period
effectively eliminate congressional oversight of these
decisions?
Fifth, how important are high-performance computer controls
to problematic Tier III countries, such as China, to the U.S.
computer industry? Does requiring licenses for these sales hurt
our industry, given that the major industrialized countries are
subject to no high-performance computer licensing requirements
and most other countries are subject to restrictions only at
much higher levels of computing power?
Sixth and finally, what affect would institutionalizing the
concepts of foreign availability and mass-market decontrols
have upon other controlled technologies? What additional
technologies would we have to make available without a license
if we wrote these criteria into our export control laws?
This Committee has been closely involved with non-
proliferation policy and export controls issues for many years.
Senator Cochran's Subcommittee has also done excellent work in
this field in recent years. I look forward to hearing our
testimony from four distinguished witnesses today who can help
shed light on these and related export control issues as we
continue our Committee's involvement with these important
national security matters in the future.
Senator Lieberman.
OPENING STATEMENT OF SENATOR LIEBERMAN
Senator Lieberman. Thanks very much, Mr. Chairman, for
calling this hearing today, which is another in a series that
we have been holding over the past couple of years on the
subject of export controls. As you have indicated, at issue
here is how we in the Congress can balance our desire for the
high-tech industry to remain healthy and robust against the
risks of allowing potential adversaries access to technologies
they may use against us.
Every time we visit this issue, I am struck, as I know most
of us are, by the paradoxes of the age in which we live, which
are, I suppose, common to all ages of innovation, but
particularly this one, where innovation is occurring so broadly
and rapidly. On the one hand, technological innovation has
significantly improved our lives, of course, by revolutionizing
how we communicate and how we live, speeding up the transaction
of business and broadening the information, education and
entertain options available to us and our children.
Innovations that most of us could not literally have
conceived of a generation ago have fundamentally changed our
lives, and we are now so immune or so perhaps numbed by the
pace of change that the remarkable and stunning very quickly
becomes commonplace, even taken for granted.
On the other hand, the precise factors that have improved
our lives have also exposed us to new dangers, because however
great technology's promise for good, the risk that it will be
used for harm is also great, and we have seen this powerfully
and painfully in the century just concluded.
This dichotomy is manifestly apparent with respect to high-
performance computers. Levels of performance once powerful
enough to qualify a product as a supercomputer now reside on
top of our desks, indeed, in our children's PLAY STATIONS. Yet
the same power that has transformed our daily lives for the
better also has potentially dangerous military applications.
So dealing with this paradox in the context of export
controls on high-performance computers is particularly
complicated and important, because not only does the technology
change at head-spinning speeds, it is disseminated at head-
spinning speeds. The difficult question before us in Congress
this year, posed specifically in the legislation that Senator
Thompson referred to, is whether we can find the appropriate
balance between economic and technological dynamism and
national security.
Mr. Chairman, these are extraordinarily important
questions. They are of great consequence to our lives, to our
living, and to our national security, and I look forward to
hearing today's witnesses and talking with them in the hope
that we can shed some light on these issues.
I should say by way of disclaimer at the outset, protecting
one of the witnesses, that I do not now, nor have I, to the
best of our knowledge, ever been related to Robert Lieberman.
Chairman Thompson. I am sure he appreciates that statement.
[Laughter.]
Senator Lieberman. He does. I wanted to clear his name here
at the outset. He is not accountable for anything I have said
or may say here today, but I welcome him and the other
witnesses.
Chairman Thompson. Well, we welcome all of our witnesses
today. We have some excellent ones: Harold Johnson, Associate
Director, International Relations and Trade Issues, National
Security and International Affairs Division of the GAO.
Mr. Johnson, do you have a statement you care to make?
TESTIMONY OF HAROLD J. JOHNSON,\1\ ASSOCIATE DIRECTOR,
INTERNATIONAL RELATIONS AND TRADE ISSUES, NATIONAL SECURITY AND
INTERNATIONAL AFFAIRS DIVISION, U.S. GENERAL ACCOUNTING OFFICE
Mr. Johnson. Yes, I do, sir.
---------------------------------------------------------------------------
\1\ The prepared statement of Mr. Johnson with an attachment
appears in the Appendix on page 40.
---------------------------------------------------------------------------
I am pleased to be here today to participate in this
hearing. My testimony is based on work that we have completed
over the last approximately 3 years. We have issued several
reports. We are currently doing some work for the Senate Armed
Services Committee. That work is in progress, so for the most
part, I will not be discussing that, but will rely on work that
we have completed.
You have my prepared statement, so I would like to
summarize just briefly a few points that are included in that.
One deals with our concern that the Executive Branch has not
fully assessed the national security risk associated with the
export of high-performance computers. Second, I want to talk
about how the Executive Branch has determined that export of
computers at existing performance levels can no longer be
controlled. And, finally, a few observations on post-shipment
verification.
Both you and Senator Lieberman have mentioned the balance
that we attempt to strike between our commercial interests and
our national security interests, and I will not comment further
on that, although one of the underlying problems that we see in
trying to achieve this balance and manage the risks associated
with export of high-performance computers is that the Executive
Branch really has not clearly articulated the specific national
security interest it is trying to protect at various computer
performance levels, nor has it stated how countries of concern
could benefit from using such computers.
We believe that without a clear analysis and explanation of
the national security interest in controlling the export of
high-performance computers, the U.S. Government really cannot
determine what militarily critical computer applications need
to be controlled or, second, the most effective way of
implementing such export controls.
If such an analysis was made, it might lead to a conclusion
that the current reliance on MTOPS as the sole measure of
computer sensitivity would no longer be appropriate. Indeed,
with the rapid changes in computing architecture and the growth
of what is called distributed computing, new approaches may be
necessary to protect the national security interests in
limiting potential adversaries' use of these machines in their
research and development programs and in their deployed weapons
systems.
In this regard, our September, 1998 report had recommended
that the Secretary of Defense make such an assessment of the
national security threat and proliferation impacts of high-
performance computers to countries of national security
concern. We thought that, at a minimum, the assessment should
state how and at what performance levels countries of concern
use computers for military modernization and proliferation
activities, and second, what impact such uses have on our
national security interests.
I would like to point out that a critical analysis of the
national security applications of concern may lead to
conclusions that are very different regarding the export
control levels that are currently in place. Indeed, the
Executive Branch may conclude that significant national
security concerns involve computer performance levels that are
at even higher levels than are currently controlled, but that
analysis simply has not been done, so we do not know that.
Despite not having done the national security analysis, the
Executive Branch has relaxed export controls on computers four
times since 1993 because it believed that machines at the
previously approved levels had become so widely available on
the market that their export was uncontrollable, and we fully
acknowledge that the computer technology has grown
exponentially. There is no doubt about that.
However, it has not been possible for us to adequately
assess the administration's justification for relaxing high-
performance computer control levels, because the term ``widely
available'' and ``uncontrollable,'' used in explaining the
policy change, has not been defined. Commerce has recently
defined controllability, and that definition includes the
criterion of volume of sales. Nonetheless, the Executive Branch
has relaxed controls based on anticipated, not actual, sales.
The Executive Branch established new computer control
thresholds based on the technical performance ratings of those
processors the computer manufacturers said would be in their
next mass-produced processor and on the estimated dates that
they would be introduced in the market, rather than on actual
volume of sales.
For example, the control levels for Tier III countries
announced by the President in July of last year roughly match
the expected performance levels of computers using four and
eight Intel Pentium processors that are expected to be on the
market in July of this year. Last November, we reported on
changes in export computer control levels the President
announced in July. We found that the administration's
conclusion was correct, that the capabilities of high-
performance computers and related components, from both
domestic and foreign sources, are generally increasing.
This conclusion was supported by evidence that they
presented in a report; however, it was true in part because the
United States does not generally control the export of computer
processor components. Most sources of this supply are U.S.
companies.
Our earlier 1998 review showed that subsidiaries of U.S.
computer manufacturers dominate the overseas high-performance
computer market, but they must comply with U.S. controls. A
1998 study, sponsored by DOD and Commerce, similarly found that
the United States had dominated the international computer
market, at least in the mid- and high-range performance
categories.
Under current regulations, computer processors that perform
up to 3,500 MTOPS can be directly exported to civilian end
users in many Tier III countries, including China and Russia,
and exports of these processors to users in other Tier III
countries, such as Israel and Saudi Arabia, are not subject to
any MTOPS levels that require a license.
Exports of other key components for systems with four and
eight processors are also generally not controlled, and these
parts can be shipped to Tier III countries for civilian end
users who can then use them to assemble computers.
Just a brief comment on the government's end-use monitoring
through post-shipment verifications. While post-shipment
verifications are important in detecting and deterring physical
diversion of computers, they simply do not verify computer end
use.
According to Department of Energy officials, it is quite
easy to conceal how a computer is being used, and although it
is possible to verify how a computer is being used by reviewing
the internal operations of computer data, this is very costly
and intrusive and requires some very sophisticated computer
analysis.
With that, I will conclude my summary and be prepared to
respond.
Chairman Thompson. Thank you very much. Our next witness
will be Robert Lieberman, Assistant Inspector General for
Audits, U.S. Department of Defense.
Mr. Lieberman.
TESTIMONY OF ROBERT J. LIEBERMAN,\1\ ASSISTANT INSPECTOR
GENERAL FOR AUDITING, OFFICE OF THE INSPECTOR GENERAL, U.S.
DEPARTMENT OF DEFENSE
Mr. Lieberman. Thank you, Mr. Chairman. I appreciate the
opportunity to be here this morning. In my written statement, I
have attempted to recap the most important findings from recent
IG reviews of the export control processes. Now I would like to
highlight four factors that my office believes merit
consideration in terms of new dual-use export control
legislation. I want to emphasize that these views are ours and
do not necessarily reflect those of other IGs or DOD managers.
---------------------------------------------------------------------------
\1\ The prepared statement of Mr. Lieberman appears in the Appendix
on page 57.
---------------------------------------------------------------------------
First, we believe that the Export Administration Act, which
expired in 1994, needs to be reenacted, rather than having the
government continue to operate under the current patchwork of
emergency declarations, other laws and executive orders.
However, any legislation in this area is going to send very
strong signals to every exporting country in the world, so it
is imperative that the law be well-thought out and the entire
spectrum of views here be carefully considered.
Second, it is vitally important that the export license
review process be properly applied. By this, I mean that it
should not be easily circumvented. In accordance with that
precept, I urge particular attention to formulation of the
control list, commodity classification requests, determinations
of foreign availability or mass-market status, and other issues
bearing on licensing exemptions. I will return to those
specific points in a moment.
Third, we believe that all available government expertise
must be brought to bear on export license application reviews.
Therefore, the current requirement in Executive Order 12981 for
Commerce to refer all dual-use license applications to Defense
for review should be made a matter of law by including it in a
new EAA.
Likewise, the exporter appeal process on licensing
decisions should be formalized in a new EAA and provide for
participation by all interested agencies.
Fourth, no program will be credible unless there are viable
inter-agency dispute resolution procedures with final
adjudication by the President, if necessary. We believe it is
particularly important to provide statutory underpinning to the
inter-agency dispute resolution process.
Furthermore, we strongly recommend that a new EAA specify
that this process be applicable to all inter-agency export
control issues, including the composition of the control list,
commodity classification determinations, licensing exemptions,
etc.
Returning, if I may, to safeguards against circumventing
the licensing process, I would like to underscore our
conclusion that the current process, wherein a DOD-developed
list of militarily critical technologies is integrated into the
overall control list, is working reasonably well.
No official, except the President, should be able to
override the determination of the Secretary of Defense that an
item belongs on the control list. Similarly, it is important
for the national security community to be involved in the
commodity classification process, which matches a prospective
export item with an export control classification number.
Those determinations are extremely important because they
indicate whether an item requires an export license for a given
destination, and if so, whether it is licensable by Commerce or
State. On pages 14 through 16 of my written statement, I
describe the joint IG review finding from last year that
Commerce was referring very few commodity classification cases
to Defense for review. This was occurring despite current
policy that requires Commerce to share with State and Defense
all commodity classification requests for items or technologies
specifically designed, developed, configured, adapted and
modified for military application.
Of thousands of requests received in a recent 3-year
period, only 12 were referred by Commerce to Defense for
review. This is an issue that actually bears on export controls
for both dual-use and munitions items. Similarly, I would like
to emphasize the need for careful controls over any process for
exempting items from licensing requirements because of foreign
availability or mass-market status.
Again, we believe that no determination to exclude or drop
an item from the control list should be possible without
Defense concurrence, unless the President directs otherwise. We
would not support any proposed legislative or regulatory
language that would allow, for example, items that would help
proliferate weapons of mass destruction to be exported without
export licenses, merely on grounds that similar items are
available from other sources.
Finally, I think it bears noting that dual-use export
license applications made up only 22 percent of the nearly
58,000 applications for export licenses received by the Federal
Government last year. Most applications go through the
munitions control process, and that is where the most concern
about excessive delay and red tape appears to have been
warranted.
That concludes my statement, Mr. Chairman. Thank you again
for considering our views.
Chairman Thompson. Well, thank you very much.
Our next witness is Gary Milhollin, Director of the
Wisconsin Project on Nuclear Arms Control. Thank you for being
with us.
TESTIMONY OF GARY MILHOLLIN,\1\ DIRECTOR, WISCONSIN PROJECT ON
NUCLEAR ARMS CONTROL
Mr. Milhollin. I am pleased to testify before this
distinguished Committee, Senator. In my written statement, I
have requested that three items be submitted for the record.
They are articles that I have recently written. I assume there
is no objection to that.
---------------------------------------------------------------------------
\1\ The prepared statement of Mr. Milhollin with attached articles
appear in the Appendix on page 76.
---------------------------------------------------------------------------
Chairman Thompson. It will be made a part of the record,
without objection.
Mr. Milhollin. You have requested that I discuss foreign
availability and mass-market status, in particular you have
asked whether these concepts are useful or not for use in
export control implementation or policy.
In order to respond to that question, I took five items
that have been controlled for some time by the United States
and our allies, and I compared the criteria for mass-market
status and for foreign availability status to those items, and
I have indicated in my written statement how that turns out.
I believe that all five of these rather sensitive things
would be decontrolled under the sweeping language that this
bill contains. I am not sure that its drafters intended for
this result to occur, but it is of great interest to compare
the criteria, for example, to high-precision switches--these
are switches that are incorporated into a nuclear weapon firing
circuit. Recently, Saddam Hussein tried to obtain 120 of these
switches as spare parts for kidney treatment machines.
He did not get them, at least according to Siemens, from
whom he ordered them, but he certainly tried. And so, what that
shows you is countries like Iraq are still trying hard to
procure items that are controlled, and I think if you look at
the criteria and you compare it to high-precision switches,
these switches would fit that definition. The bill says that if
they do, the Secretary of Commerce must decontrol them. It
gives him no discretion, and the same is true of many other
items that we have controlled for a long time.
Glass and carbon fibers are another example; these are used
to make the rotors for centrifuges to enrich uranium for
nuclear weapons; they are used for rocket cases; they are used
for rocket nozzles; they are used for rocket nose cones; they
have been controlled for a long time. We apprehended a person
here who was trying to send this material to Cairo for use in a
missile that the Iraqis were never successful in building. They
were building it in cooperation with the Egyptians and the
Argentines.
If that had succeeded, then when we deployed our troops in
the Gulf War, it would have been a very different scenario. If
Iraq had had the kind of missile that it was building with
these fibers, history would have been different. So we are not
talking just in theory here about dollars and cents. We are
talking about actual threats to our troops.
Maraging steel is another item I looked at. Maraging steel,
as well, it was one of the few materials that can make high-
speed rotors for centrifuges to enrich uranium. Maraging steel
is also used in missile applications. We have protected it for
a long time. We apprehended a Pakistani who was trying to
export this steel to Pakistan some time ago. It, too, in my
judgment, would be caught by the sweeping language of this
bill.
The other things that I looked at were corrosion-resistant
valves. Those are used to resist the corrosive material in
uranium enrichment plants. Iraq and Iran, when they go the last
step in building a uranium enrichment plant for nuclear
weapons, will need lots of these valves. You cannot build a
plant without them. That is why they are controlled for export.
If this language passed in its present form, I think these
would be decontrolled. High-performance computers also would
fall under this category, and the reason for the presence in
the bill of these concepts is because the computer industry has
pushed for them to be included. I think it would be a very
dangerous thing to decontrol high-performance computers just
because they are made in large quantities domestically.
We have always used--the United States has always used its
highest-performance computers for designing nuclear weapons and
for cryptography. It is reasonable to expect other countries to
do the same. The Russians, after they illegally imported
supercomputers from us--that is, from IBM and Silicon
Graphics--announced that they were planning to use those
computers to design nuclear weapons after the test ban came
into effect--that is, the present moratorium on testing.
So we know these items have great national security
significance. To decontrol them under a vague criterion, such
as mass-market status, in my opinion, would be a big mistake.
The other concept that you have asked me to discuss is
foreign availability. That, too, would, I think, decontrol many
things that its drafters did not intend to decontrol. Just for
purposes of illustration, I compared North Korean rocket motors
to the criteria in the bill for foreign availability. If you
look at those criteria, I think you will see that, actually,
North Korean rocket motors would be decontrolled; that is, they
are foreign-available under the definition in the bill.
The criteria say that for something to have foreign
availability status, it must be available to controlled
countries from sources outside the United States. North Korean
rocket motors obviously are. Lots of countries are buying them.
Also under the criteria, the motors can be acquired at a price
that is not excessive and they are available in sufficient
quantities so that the requirement of a license or other
authorization with respect to the export of such item is or
would be ineffective.
I do not think the drafters of this bill intended to
decontrol motors, but it looks to me as if they have, and they
have swept in a lot of other things, as well. I do not think
that the definition of foreign availability, as now written
into the bill, is really tolerable. It would require the United
States to decontrol things that our allies control under
regimes that the United States has built up--the missile
technology control regime; nuclear suppliers group guidelines--
all the regimes that have been built through U.S. diplomacy
since World War II.
If we apply this language literally to the things now on
the list of those regimes, our government would be required to
decontrol a great many of them. This would leave our allies
aghast, and it would--well, I do not want to go so far as to
say it would make us into a rogue supplier, but it would
certainly break our international obligations and it would give
a signal to the rest of the world that we really did not care
about export controls.
I think the reason that the bill is so sweeping is because
it has adopted a principle that really is not sustainable, and
that principle is that if somebody else is doing it, we should
do it, too. The United States has never followed that kind of a
principle in our own actions or in our diplomacy toward other
countries.
I had the dubious honor of being on CNN a lot during the
Gulf War, and testifying before Congress about scuds and about
other things that turned out to be a big surprise to the world.
One of the things I remember was the Israelis holding up pieces
of scuds that they had found in the debris of destroyed
buildings in Tel Aviv. They found German markings on some of
those scud fragments.
If you adopt the idea that our industry should be able to
sell anything any other industry should sell, then you have to
accept the idea that somebody is going to hold up a fragment
with a U.S. marking on it. I cannot believe our industry really
wants that to happen.
We also should remember that a scud supplied by Russia and
enhanced in range by Germany killed our troops in Saudi Arabia,
sleeping in their barracks. I cannot believe that the United
States would want our industry to be able to participate in the
market that caused that to happen, even if it means losing
sales and even if it means that the countries like Iraq, who
are doing these things, can get it from somebody else. You
simply cannot go down the road in which you say, ``If somebody
else is going to do it, our guys should be able to do it,
too,'' unless you are prepared to sustain the criticism and the
shame that would result from seeing your products used to
achieve the things that were achieved by Saddam Hussein.
So I think that is the main problem we are getting into, is
that we have these arguments that if somebody else is going to
do it, we should do it. That is not a position you want to
take. It would make export controls, as a practical matter,
impossible in the world if everybody adopted that point of
view.
The last thing I would like to mention here is that in the
latter portion of my testimony--other portions of my testimony
refer to things that have been covered by other witnesses--in
the last section of my testimony, I recommended that one of the
ways around the dilemma that Senator Lieberman mentioned would
be simply to use transparency; that is, if we put more light on
the export control process, I think it would cause us to do a
better job. It would allow our government to deny things that
are dangerous and allow legitimate trade to go forward.
I have attached to my testimony a proposed list that is a
first step toward more transparency. I have attached a list of
50 Chinese companies that I believe, based on very reliable
open-source information, are dangerous buyers for high-speed
computers and for other dual-use technologies. I recommend that
the Committee submit those names to the State Department for
review; and if the State Department agrees that they are
dangerous buyers, then they should be put on the warning list
to U.S. exporters.
I must say that I am glad that Dan Hoydysh is here today,
because when I interviewed him for an article in The Washington
Post that I wrote not too long ago about this subject, he
agreed that the industry would welcome more guidance on who the
bad guys are, who are dangerous buyers. I am not suggesting
this list as an embargo list. I am suggesting it simply as a
warning list; that if an exporter gets an order from one of
these buyers, the exporter should check it out.
What I am recommending is that it should trigger an export
license application. I am not saying that these recipients
should be denied U.S. exports. I am just saying that it merits
a look if they are going to buy something that can conceivable
contribute to a nuclear weapon or missile program.
Thank you very much.
Chairman Thompson. Thank you very much.
Our last witness will be Daniel Hoydysh, Co-Chair of the
Computer Coalition for Responsible Exports.
Mr. Hoydysh.
TESTIMONY OF DAN HOYDYSH,\1\ CO-CHAIR, COMPUTER COALITION FOR
RESPONSIBLE EXPORTS
Mr. Hoydysh. Thank you very much, Mr. Chairman, for the
opportunity to testify on this important subject. I have
submitted my full testimony and in the interest of time I would
briefly summarize a few highlights. Before I get to that, I
would like to make two points in response to what my good
friend Gary has said.
---------------------------------------------------------------------------
\1\ The prepared statement of Mr. Hoydysh appears in the Appendix
on page 102.
---------------------------------------------------------------------------
One, I believe that in the Export Administration Act, no
decontrol is automatic. There are provisions for presidential
overrides, and that is something that at least ought to be
looked at carefully. Two, we do not advocate decontrolling
computer exports to any of the rogue or terrorist nations.
I would like to make a couple of other major points.
Something often gets lost in the heat of debate is that our
industry cares very deeply about national security. We are
responsible citizens of the United States and would not do
anything to jeopardize the security of the country in which our
workers and families live. We believe, however, that our
national security is directly dependent on the technological
leadership of the U.S. computer industry.
To maintain this leadership, we must compete in the global
market and we must export. Exports equal profits. Profits are
used to fund R&D, and R&D drives technological leadership. The
U.S. computer industry is the world leader and we want it to
stay that way. But we do have substantial foreign competition.
If I accomplish nothing else in this testimony, I would like to
lay to rest the myth that there is no foreign competition for
the business computers that we are talking about in terms of
decontrol.
According to a recent report by the International Data
Corporation, four of the top 10 server vendors are foreign.
They include companies from Japan, Germany, and France. We are
not proposing that controls on supercomputers be abolished.
Again, I would like to dispel the myth that this debate, is
about the export of supercomputers.
We are proposing that restrictions be eased on business
servers with two, four, and eight processors. These machines
are the basic building blocks of the new digital Internet
economy.
Supercomputers, such as those used for sophisticated
nuclear simulations, consist of thousands of processors. For
example, in the Sandia Labs, Intel has installed a machine
called the ASCII-RED, which has 9,632 processors. Recently, the
French Atomic Energy Commission ordered a supercomputer from
Compaq for simulation programs to ensure the reliability and
safety of the French nuclear stockpile without the need to
conduct new nuclear tests. The Compaq system will use 2,500
alpha processors, will take a year to install and operate
roughly at 5 million MTOPS. Now that is a supercomputer.
Finally, Fujitsu recently announced that it would provide
the world's most powerful supercomputer to the Toyota
Corporation for automobile design purposes. In its maximum
configuration, this system has 512 vector processors and can
operate at over five million MTOPS. And please note that this
system is replacing a U.S.-made Cray supercomputer.
While we do not come to praise the export control system,
we certainly do not come here to bury it. We support effective
export controls. However, we think the current system is
broken. It is broken because it is inefficient, it is
ineffective, and it is counterproductive.
It is inefficient because it takes too long to process
export licenses. The reviewing agencies do not have the
resources, either in personnel or equipment, to do the job.
It is ineffective because it is largely unilateral. Our
controls are much stricter than those of our foreign
competitors. It is counterproductive because it wastes
government and industry resources in trying to control the
uncontrollable. Therefore, efforts to police truly sensitive
items are diluted, and this undermines national security.
Dr. Steve Bryan, a respected expert on export control, who
served in the Reagan Administration and is in large measure the
architect of a lot of the systems that we have, described the
current export control system best when he testified before the
House Armed Services Committee last year. And I quote: ``I do
not think there is any point in having an export control system
that tilts at windmills. I think you have to have controls that
makes sense, that can be enforced, and that protect our
strategic interest.''
Another quote that is right on point in terms of evaluating
the export control system is contained in this report which I
recommend to the Committee, ``Final Report of the Defense
Science Board and Task Force on Globalization and Security.''
The basic premise of this report is that rather than trying to
restrict exports of widely available technology, we have to
concentrate on trying to run faster than our adversaries.
And a quote that is particularly on point here from this
report, ``Protection of capabilities and technologies readily
available on the world market is, at best, unhelpful to the
maintenance of military dominance and, at worst,
counterproductive by undermining the industry upon which U.S.
military technological supremacy depends.''
So what do we need? We need to fix the export control
system. We need an efficient, effective, and credible control
system that reflects competitive and technological reality. In
the short-term, we would like to reduce the congressional
review period from 6 months to 30 days.
We also need to increase control thresholds now to reflect
advances in technology and competitive reality. In the long-
term, we want to work with the Executive Branch and the
Congress to develop an effective approach to controlling
exports that fit national security concerns of the 21st
Century. This would require a thorough evaluation of the
threats posed to the United States in this century, the effect
of globalization of markets technology and knowledge,
identification of choke-point technologies and techniques for
how we can run faster than our potential adversaries.
Thank you very much.
Chairman Thompson. Well, thank you very much. This
testimony, from all of you, really lays excellent groundwork
for our discussion today, and it is a classic case of two
competing interests. Both of them are valid interests, but I
was just thinking yesterday about what was happening here in
the process, as we are trying to balance our commercial
interest and our competitiveness--and keeping our own
capabilities where we want them to be--versus national
security.
It seems to me all the movement--we can argue about this
later--but it seems to me all the movement seems to be on the
side of the export industry. The administration, of course,
periodically and quite often increases the MTOPS levels at
which computers can be sold without a license. They were
changed in April, 1994; October, 1995; July, 1999; February,
2000; and again in August, 2000, as I understand it. So we are
moving in that direction.
We have gone from 2,000 MTOPS for military use in October
1995 to conceivably as much as 40,000--according to what some
industry folks were saying earlier this year--in August 2000.
You might ask, compared to what, because some computers have
MTOPS levels in the millions, but we are moving certainly in
that direction. Congress has a right to review whether or not
we are moving too fast or perhaps not fast enough. Up until, I
believe, 1999, Congress had what was in practice an 18-to-24
months review period. So until recently, we had 2 years. Now we
have got 6 months.
Now it is being suggested by proposed legislation that we
reduce that review level to 30 days. I think it is important to
keep in mind this has nothing to do with the holding up of an
individual export. It just has to do with whether or not
exports at particular levels even need licenses at all. So that
may be reduced.
Then we have the Export Administration Act, which
interjects new concepts in terms of statutory law, derived and
greatly expanded from what previously were only to be used by
exporters who were denied licenses and so forth. Basically, the
concepts and statutory law of ``mass-marketing'' and ``foreign
availability'' are new, and this bill would propose to take
whole categories of items, even above the MTOPS levels that are
allowed, out of the control regime completely on the basis that
everybody supposedly has got or can get them anyway.
So there is quite a bit happening here, and it seems to me
that it is all moving in the same direction. Now, perhaps the
case could be made that this is good and that it is valid. If
we err too far on one side, we are perhaps hurting ourselves
somewhat commercially. If we err too far on the other side, we
are perhaps hurting ourselves somewhat from a national security
standpoint.
And one of the things that concerns me from your testimony,
Mr. Johnson, it that as the administration makes this
determination as to when and how much to raise these MTOPS
levels, they are not making any kind of national security
assessment. It is strictly based on what is deemed to be
controllable or uncontrollable. Is that correct?
Mr. Johnson. That is basically correct. They have not done
a national security analysis to know what impact the relaxation
or the change in the control levels might have on our national
security and how the recipient governments, may use the
computers in their military modernization programs, and we
think that that is a serious deficiency.
Chairman Thompson. And the GAO, of course, has been dealing
with this for some time, and you have had occasion to criticize
the administration in times past because of some of the
analyses and studies that they were relying upon, such as the
Stanford analysis, in making their decisions to raise the MTOPS
levels.
Mr. Johnson. Correct.
Chairman Thompson. So the justification for their decision
is one of the things that we can look at in trying to determine
where we should go and how fast we should go as an
administration is increasing these MTOPS levels. I am certainly
not arguing that they should not be increased. But the question
is to what levels and how fast? Reasonable people can disagree
on that, but one of the benchmarks that I think we can look at
is the nature of the material they are relying upon in order to
make those increases.
The fact of the matter is that in times past, in dealing
with these nebulous terms of foreign availability or
uncontrollability or whatnot, they have relied on studies that
you did not feel supported the conclusions that the
administration came up with. Is that not accurate?
Mr. Johnson. Well, it was unclear whether the studies
really supported the conclusion, because the studies themselves
lacked empirical data to support the conclusion that was in the
study. And mainly in the area of controllability, the study
simply did not have sufficient data to come to a conclusion
that--I think the initial Stanford study indicated that
computers at 4,000 to 5,000 MTOPS were, at that point in time,
which was 1995, uncontrollable.
There simply was not data to support that, so whether or
not they came to the right conclusion, we did not reach that
conclusion. They may have serendipitously come to that
conclusion properly----
Chairman Thompson. It would seem to me, that there's so
much anecdotal evidence--so many statements that we hear from
time to time about the clear availability of computers when you
walk into Radio Shack, etc. If that is so clear, you think they
would be able to come up with a study that the GAO would say at
least is a valid study in order to support that conclusion.
Mr. Johnson. Well, at certain MTOPS levels, I am sure they
could do that. But, we are talking about MTOPS levels that are
generally higher than what you would find at Best Buy or Radio
Shack.
Chairman Thompson. And the administration deals not just in
terms of what is perceived to be the case at the time of the
analysis, but also of the anticipated availability.
Mr. Johnson. That is a major concern that we have had in
the last couple of--in our current study, as well as the study
that we did of the President's July, 1999 report; that the
decision was based on anticipated mass-market, rather than on
what existed at the time the decision was made.
Chairman Thompson. Correct me if I am wrong, but it seems
to me that this anticipated mass-market, in turn, is based upon
what our domestic producers intend to manufacture in the
future.
Mr. Johnson. That is correct, what they say they are going
to produce and when they are going to----
Chairman Thompson. So foreign availability----
Mr. Johnson. That is not a factor in that kind of judgment.
Chairman Thompson. Well, is controllability?
Mr. Johnson. Well, projected controllability, yes; but
suggesting that because they are going to have a particular
type of processor available 6 months hence does not necessarily
mean that we should be decontrolling now in anticipation that
the processor will be available.
If you are looking at what is mass-market, I mean, that is
something that exists, you can count--I do not want to put
numbers on what the criteria ought to be, but you can determine
what a mass-market is, rather than what is anticipated.
Chairman Thompson. All right. Let's make sure we are
talking about the same thing. Let's talk about not what it
ought to be or what you think it should be. Let's talk about
what the current situation is, as it is applied now, as these
determinations are made to raise these MTOPS levels. It is
based in part on anticipated levels.
Mr. Johnson. That is correct.
Chairman Thompson. And that, in turn, helps to reach a
determination as to what is going to be controllable.
Mr. Johnson. What the control levels ought to be.
Chairman Thompson. Ought to be?
Mr. Johnson. Right.
Chairman Thompson. So it sounds to me like a self-
fulfilling prophecy. If your supposed ability to control or the
controllability of a type of item is determined primarily on
the basis of what our domestic manufacturers intend to produce,
that seems to be a self-fulfilling prophecy. I mean, that begs
the question: Should they be controlled, should they be
available? And certainly we have something to do with that.
That is my assessment.
I mean, that is my comment. Do you have any problem with
that?
Mr. Johnson. I think that is a fair analysis, yes.
Chairman Thompson. All right, sir. One more thing, on the
reduction of the time of the analysis. You mentioned you were
doing some work for Armed Services.
Mr. Johnson. Right.
Chairman Thompson. Is part of that work an analysis of the
last MTOPS level proposal, I will call it, of the
administration's----
Mr. Johnson. Yes.
Chairman Thompson [continuing]. Increase?
Mr. Johnson. Yes. It is an analysis of the President's
February, 2000 announcement that----
Chairman Thompson. All right. I will just ask you, do you
have any anticipated date of release of that?
Mr. Johnson. Well, probably within 4 to 6 weeks.
Chairman Thompson. Probably within 4 to 6 weeks.
Mr. Johnson. Right.
Chairman Thompson. So the GAO does that. Do you
traditionally do that? I mean, is this your first time?
Mr. Johnson. No, we have done this twice now.
Chairman Thompson. OK.
Mr. Johnson. And it is anticipated--we have had some
discussion at the staff level that GAO might be requested to do
this on a routine basis.
Chairman Thompson. All right. Now, my information here--you
correct me if I am wrong--that from September 1993 through
October 1995, the review time period was in practice for 18 to
24 months. From July 1999 till now, it is 6 months. The
proposed legislation would cut that review time back to 30
days. As the entity that is doing that review, what is your
opinion of that?
Mr. Johnson. I think 30 days would reduce unreasonably the
amount of time the Congress has to look at the President's
report. In terms of our work, we do not require the full 6
months. I mean, we have been--we have had this study underway
now for probably 6 or 8 weeks, but if we had immediate access
to information from the Commerce Department when the
announcement is made, that would shorten our time frame.
So I am not suggesting that it needs to remain 6 months. It
can be shortened from that, possibly, but I think 30 days would
be overly-restrictive for the Congress to deal with it.
Chairman Thompson. So you have had this review underway for
about how long so far?
Mr. Johnson. Probably about 8 weeks.
Chairman Thompson. And you anticipate, you said a few
minutes ago, how many more weeks?
Mr. Johnson. Well, I would say 4 to 6 weeks, until the
report is published; but, I mean, under other circumstances,
that time could be shortened to some extent.
Chairman Thompson. All right. Now, that is just your time
to produce a report. That is not congressional review or
analysis or hearings or anything else.
Mr. Johnson. That is correct. That is our time.
Chairman Thompson. I see. All right. Well, I have taken up
more than my time.
Senator Lieberman go ahead.
Senator Lieberman. No problem. Thanks, Mr. Chairman, very
interesting discussion.
Mr. Milhollin, let me start asking you a question that goes
back to something you said and at least helps me get into this
discussion and the competing interests and values involved
here. At one point, you quoted one of the common responses to
this dilemma, which is, ``Well, if everybody else is doing it,
we should, too,'' and you were critical of that.
In one sense, of course, you are appealing to America's
better nature. We like to believe that we are not like everyone
else, both in terms of values and hopefully the extent to which
we are prepared to protect our national security interest. But
I want to just start my discussion with the panel by asking you
whether that criticism of that response, if everyone else is
doing it, we should, too, you mean it comprehensively?
In other words, if you were convinced that, in fact, some
high-performance computer was really quite widely available--
let's assume for a moment that the facts were proven--but that
it really could be used to endanger our national security,
would you still say we should not do it--and, of course, we all
remember those moments in the Gulf War when the shards of
different systems were held up--to avoid in some measure, to
having blood on our hands?
Mr. Milhollin. I would answer your question by saying first
that it is not simply a moral position I am taking, although I
think it principally is--that is most of it--but it also is a
functional point. If every member of the regime operated on the
assumption that everybody else was going to sell anything he
did not sell, it would be pretty hard to have a regime.
It is really a question of keeping the faith. Somebody has
to be the leader. It has always been the United States. We have
gone to our allies and said, ``Here is what we think ought to
be controlled.'' We did that in COCOM. We said, ``Here is what
we think is important. Here is what we think ought to be
controlled. We are going to control this and we hope that you
will join.''
In fact, that was President Bush's approach to the Gulf
War. I mean, we did not wait for everybody else to decide,
``Yes, we are going to roll this back.'' President Bush said,
``We are going to roll this back,'' and then he invited
everybody else in to come in and help out.
Senator Lieberman. Right.
Mr. Milhollin. As a practical matter, that is how you have
to do it. So it is not just a moral issue. It is also a
practical question of how you achieve things diplomatically,
and unless somebody is going to step out there and take the
lead, nothing happens.
Senator Lieberman. But what if they do not follow on these
commercial questions, on the sale of high-performance
computers?
Mr. Milhollin. They will never follow 100 percent.
Senator Lieberman. Right.
Mr. Milhollin. What you have to do is decide how much is
enough. Under COCOM, the Russians could always get things.
They could always--if they wanted something enough, they
could figure out a way to get it. But, often, they did not get
training; they did not get manuals; they did not get spare
parts; and 6 months later, it wound up being a piece of junk
because they could not service it. They admitted that after the
Cold War ended and there was sort of a look at how COCOM had
functioned.
COCOM was a giant success not because it was airtight--we
had the Toshiba case. We had lots of situations where people
violated COCOM and undermined our industries, but overall it
worked, and if you visit the former Soviet Union now, you can
see the impact it had on their infrastructure.
Go to Russia and try to make phone calls from one village
to another. COCOM really did have a big negative impact, on
Russia primarily, even though it was not airtight; that is, it
did not work 100 percent. So you will always have situations
where somebody will not follow. What I am worried about is that
we have to maintain the faith here or we will not have anybody
following. That is really the question.
It is not the question of whether you can get 100 percent
compliance. The question is, if you decide that we are going to
send the world a message that we do not care about this
anymore, then you are certain to get zero compliance.
Senator Lieberman. That is very interesting. So is it fair
for me to conclude that what you are saying is that even if a
particular dual-use item is available in foreign markets to
foreign countries, to countries that we would put in one of the
tiers that we worry about that you would say that we should
still try to control its export from here, even if that does
some damage to our high-technology companies, because the
effect will be that we will make it harder for those who would
threaten us to get hold of it; so that this is a balance and
maybe there will be some economic damage here, but when you
balance it against the national security threat, because we are
the leader, it is worth it.
Mr. Milhollin. Yes, it is worth it. You are going to take
some hits. You are going to take some losses, but nothing is
free.
Senator Lieberman. Right.
Mr. Milhollin. So you have to be willing to say to the U.S.
industry, ``You are going to lose some sales here and there
from this system, but overall, it is worth it.''
There was a recent case in which a rather sensitive Chinese
company was tried to buy a five-axis machine tool from us and
they did not get it. I must admit, because I have an activist
hat, that the deal happened to be exposed in the newspapers and
it made it much harder to approve.
Well, I was told that an European company filled that
order. I suppose that pro-export forces would say, ``Well,
there is an example. You know, you made it hard for us to
approve this. It got held up. We did not approve it and then
the Chinese got tired of waiting and they bought it from
somebody else.'' Well, I think that is a victory. If this
company is going to make missiles and military aircraft at this
plant, and they want to do it with a German or a French or a
Swiss machine tool, I think it is better than having them do it
with ours, because if we had sold that machine tool, then the
message would have been clear, because this was a very dubious
end-user. The message would have been, ``Look, guys, if
America--if we are going to do it, then the signal is, we
really do not care that much.''
Senator Lieberman. So, your reasons are both moral and
functional or practical?
Mr. Milhollin. Yes.
Senator Lieberman. Both, we do not want that on our hands,
but also that, so long as we exercise some restraint, even
though in that case that Chinese company got what it wanted,
that we send out a message to the world that makes it less
likely that this stuff will be more widely available.
Mr. Milhollin. That is right. The next time the Iranians
want something and we go to a European company and say, ``We
discovered that the Iranians are about to buy this from you and
we think you ought to stop it,'' they are not going to say,
``Yes, but what about the machine tool you guys just sold to
the Chinese?''
Senator Lieberman. Right.
Mr. Milhollin. And, believe me, that happens. It happens
all the time. So unless you are willing to be clean in your own
behavior, you are not going to get anywhere with anybody else.
Senator Lieberman. Mr. Lieberman, from a Defense Department
perspective--I know you are not here to speak for the Defense
Department--but from within the purview that you have, how do
you react to that standard that Mr. Milhollin establishes? I
know from your testimony that you obviously feel that the
Department of Defense should be involved more in these export
control decisions, but he is posing a tough standard, I think.
I am interested in your reaction to it.
Mr. Lieberman. I believe that it serves no purpose to have
any export control regime if we do not enter into that process
with expectations of having a pretty tough standard. I would
point out that when a buyer decides to purchase a product, they
make their choice based on what they think the best product on
the market is and perhaps best price.
So what we seem to be talking about here is selling
products that could create national security threats, and we
are providing the best product at the best price to whomever
that other party is, which, to me, creates a problem. You used
the word balance in your opening statement, and clearly that is
what we are talking about here.
I believe that the current process is somewhat inefficient,
but the inefficiencies in the process can be fixed, because
those are mostly bureaucratic procedures and resourcing
questions that can be addressed.
Senator Lieberman. In some of the delay involved.
Mr. Lieberman. Yes. Absolutely.
Senator Lieberman. In reaching a decision.
Mr. Lieberman. But the basic feature of the existing review
process, which is multi-agency deliberation over the wisdom of
granting a license, I think is really the thing that we need to
retain. That is why I said that we need to be extremely careful
about exempting products or classes of products from the
process, because I think ultimately the process needs to be
applied to anything where there is even a smidgen of a national
security implication.
That does not mean when something has to go through the
process, that a license is going to be denied. In fact, very
few licenses are denied. So I think that is an important
distinction that I would try to make.
Senator Lieberman. I would take it that you believe the
bill before the Senate, S. 1712, is somewhat imbalanced, that
it tilts too much toward the commercial interest.
Mr. Lieberman. Yes. That has been our testimony to the
Senate Armed Services Committee and there is some of that in my
statement today.
Senator Lieberman. Mr. Hoydysh, how about Mr. Milhollin's
standard? I presume you find it too stringent and, in some
senses, unrealistic.
Mr. Hoydysh. Well, Senator, before I get to that, could I
just take one point to deal with a factual issue that Senator
Thompson raised earlier about the review period going from 180
days to 30 days?
Senator Lieberman. Yes. Sure.
Mr. Hoydysh. Senator Thompson mentioned something about an
18-month or 2-year review period. In my knowledge, there was no
review period before the current 180-day review was enacted in
1997. I spent 7 years in the Reagan Administration (from 1982
to 1989), in the Export Control Bureau, and there was no
congressional review period whatsoever. So I am not sure what
it is factually that the Chairman is referring to.
Chairman Thompson. I think the difference has to do with
the policy and the law. As I look at this, my chart indicates
that the 18 to 24 months was the policy during that period of
time. The 6 months was put into law, so----
Mr. Hoydysh. Fair enough.
Chairman Thompson. Whatever conclusions we want to draw
from that, I think that is what the situation was.
Mr. Hoydysh. To get back to the standard, Gary raises a
very difficult question. Moral questions are always very
difficult. I think the answer is in how you draw the balance
and where you actually draw the line. We could all agree that
no one wants to sell items to someone who is going to do damage
to this country.
In fact, the system, as currently structured, has
safeguards built into it.
Senator Lieberman. Would you say that even if the item is
available from other countries? In other words, that is part of
the challenge that is posed here.
Mr. Hoydysh. Senator, we are not permitted under current
law to sell to an end-user in China or anywhere else if we have
knowledge that that end-user is going to use that equipment for
proliferation purposes.
Senator Lieberman. Right.
Mr. Hoydysh. Most of the enforcement cases that have been
handled over the years generally result from tips that are
provided by the industry. So, with a few exceptions which have
been publicized in recent years and which are under
investigation the industry, I think, has a very good record on
that.
The real question is where do you draw the line and at what
level do you have to either control things by having the
government review each export, and where do you allow industry
to make that ultimate choice?
Senator Lieberman. That is exactly the question. So, in
other words, if Mr. Milhollin has drawn the line in one place,
which is that even if a particular dual-use item is available
in foreign markets, we ought not to be selling it, if we are
reasonably confident that it is going to be badly used, not
only for moral reasons, but because we set a standard. So that
is a tough standard. Where would you draw the line yourself?
Mr. Hoydysh. I would not remove the requirement that if
there was knowledge that the equipment was going to be used for
some inappropriate purpose, that we should not sell.
Senator Lieberman. Forgive me for interrupting, but how do
we enforce that? That is a reasonable standard, but how do we
make it real?
Mr. Hoydysh. Well, in some cases, it is easier to enforce.
I mean, most of the companies--most of our companies that deal
in countries like China, for example, have a fairly complex
process in terms of searching for customers. Generally, we just
do not go off and sell someone 50 boxes of something and leave
it there. Generally, it is involved in providing an airline
reservation system or some kind of banking system.
So we have a good idea of what the end-use is. If somebody
comes to us and says, ``I want you to help me automate the
missile factory,'' we will not do it. In fact, we are not
permitted to do it. So there is just, in the normal way of
doing business, a certain amount of security in how these
products are used. Where it begins to break down is if you have
a product like the Macintosh PowerG3 or whatever the
nomenclature is, and you want to sell that in a Radio Shack or
a Best Buy-type environment in China, where you provide 50 of
these to people who walk in off the street, as a practical
matter, it becomes almost impossible to monitor where those
sales are going to go to.
There is a big difference between delivering 150 boxes to
the People's Liberation Army missile base and delivering them
to Best Buy. So, I mean, these are just practical issues and,
at some point, it simply becomes impossible to actually do it
in practice, no matter how well-intentioned the idea is.
Senator Lieberman. In your testimony, you talked, and we
all have, about the extraordinary advances, rapid advances in
the capability of high-performance computers. Now, as you
mentioned, we are into millions of MTOPS. But isn't it true
that some of the applications that we have made that are most
significant of high-performance computers have been done at
much lower MTOPS levels, some under the limits we are now
establishing?
The question is, is the MTOPS number an adequate and
appropriate standard to use?
Mr. Hoydysh. It is becoming clear to people in industry and
people at the Defense Department, as well, that MTOPS alone may
not be an adequate standard to measure the strategic
significance of a computer.
The computers that we are talking about are basically
designed for transaction processing; that is why they have many
processors, so that Visa or a bank can handle 100,000 phone
calls a minute coming in. They really are not specially
designed to be used for military applications or for nuclear
weapons or any of the prohibited purposes.
That is not to say that they cannot be used for prohibited
purposes, because any computer can be used for that, but they
are really not designed primarily for that. There may be other
technological parameters, in addition to MTOPS or in place of
MTOPS, to better measure what it is that we are concerned
about. I understand that the Commerce Department Technical
Advisory Committee is looking at this and also the Defense
Department is looking at this.
Senator Lieberman. Well, that is a very important point and
obviously we would be interested in the results of that
inquiry, because that may help us come to a more practically-
effective standard.
Mr. Johnson, you said some things in your testimony I want
to just ask you to amplify a bit which interested me, which was
that the Executive Branch has not clearly stated what the
national security interest is in controlling exports of high-
performance computers, and, in a way, this may touch on the
last point of exchange with Mr. Hoydysh, that maybe MTOPS are
not the appropriate, certainly not sole standard, for
determining what should be exported. I just wonder if you could
talk about that a little more.
Mr. Johnson. Sure. Yes, we have basically come to the same
conclusion. In fact, another study that we have been requested
by the Senate Armed Services Committee to do is to follow these
studies that are ongoing within industry, as well as Department
of Defense and Commerce, on what other standards might be
appropriate, other than MTOPS, because MTOPS clearly does not
address the concern that has been raised about distributed
computing; in other words, where you can line up several
computers and tie them together and distribute the calculation
process among them.
It just does not address that issue. But in terms of our
concern that the Defense Department has not clearly addressed
the national security interest it is trying to protect,
obviously, they have looked at how computers are used in their
own processes, and have come to conclusions that there are a
number of applications that are important applications that are
required from a very low level of performance to a high level
of performance.
Our concern is that there are a number of applications,
computer applications, that are of such critical importance
that we need to do whatever is required to protect our ability
to deal with those applications, and that is what the Defense
Department has not done. They have been requested now, through
legislation, to do that.
I think the National Defense Authorization Act of last year
laid that requirement on. My understanding is that they will
have that study completed in August, and hopefully that will
resolve that concern; but what we have held is that if they
looked at all the applications that are of extreme importance,
they may come to a conclusion that the level that we are trying
to control computers at now just does not make any sense. It
may be much higher than what we are trying to control. It may
be lower, too, but----
Senator Lieberman. It may be lower, if we add in other
standards, other factors to consider.
Mr. Johnson. Yes. Right.
Senator Lieberman. Could you give us an example of one or
two other factors that, to you at this point, seem relevant
besides the MTOPS standard?
Mr. Johnson. Well, just basically, looking at the
architecture of the computer, what it is designed to perform,
and I am not a technician myself, but what the computer is
designed to--the application is designed to perform--would be
one way of measuring it. I know it is fraught with all kinds of
problems and we have already had some discussions with industry
on that, but----
Senator Lieberman. Mr. Johnson, a final question. You also
said that you thought that two very central terms to this
discussion, particularly of the legislative proposal before
Congress, are not adequately defined; that is, the terms widely
available and uncontrollable.
Mr. Johnson. Controllability, yes.
Senator Lieberman. Controllability. So tell us what is
lacking and, if you had your druthers--you were drafting--how
would you define those two critical terms?
Mr. Johnson. Well, I basically--the Commerce Department, in
response to our raising that issue, did define controllability,
and I think their definition of controllability is not a bad
standard. Unfortunately, they did not apply that standard. But
in defining controllability, they included factors like the
volume of sales, and I think that is a critical aspect.
I do not want to try and attach numbers to it. It is very
difficult to do that, and it would differ depending on what the
component is. But the way Commerce defined controllability to
us, which implies widely available, it is not a bad standard,
and I think we have included that statement in our prepared
testimony.
Senator Lieberman. OK. Thanks, Mr. Chairman. I am going to
give it back to you.
Chairman Thompson. Thank you very much. We have been
looking at the front end of the process as we increase the
MTOPS levels and we propose to shorten the amount of time
Congress has to review, but we have also touched on the back
end of the process--and that is, who winds up with these so-
called supercomputers?
Everything we are doing is based on an assumption that we
have something to do about that or that we can in some way
affect that or control that. In our tiered process, for Tier
III countries such as China, we have an export-license free
computer performance level for military use and one for
civilian use. But, in listening to you, I was reminded of a
point that the staff had made earlier to me, and that is about
the difficulty in dealing with a country like China, and being
able to rely upon the proper end-use of an export to a civilian
company with assurances that it will not have military use.
There is also a Russian angle to this problem. In 1996,
both Silicon Graphics and IBM illegally exported high-
performance computers to Russian nuclear weapons laboratories
without licenses. They claimed that they had not known that
these facilities were weapons labs, even though the two
locations, Chelyabinsk 70 and Arzamas-16--which I have been to,
by the way--should have been well-known to anyone with any
knowledge of the Russian nuclear program. I do not know. It
seems like I knew about Arzamas a long time ago.
Anyway, after these illegal sales were revealed, the head
of the Russian program bragged that he had planned to use these
machines to design nuclear weapons. I guess it is impossible to
keep something like that from happening every once in a while.
But, again, we need to look at the process, because we are
relying upon industry to make the initial determination of
whether a Tier III end-user is civilian or military.
That might be more than industry really has the capability
of doing. It might be an unfair burden to be putting on them.
Some in industry might be tempted to hedge. In the China
situation, there was a period of time there when they were
allowing no post-shipment verification checks at all. Isn't
that true?
Mr. Johnson. That is correct; yes.
Chairman Thompson. A period of time, up until the agreement
in, what, 1998?
Mr. Johnson. June 1998.
Chairman Thompson. June 1998. They would not allow us----
Mr. Johnson. Right.
Chairman Thompson. Here we are trying to export these
computers to them, and they would not allow us to check with
regard to post-shipment verifications. Then, we supposedly
entered into an agreement with them in 1998 that would allow
post-shipment verifications. Is that correct?
Mr. Johnson. Yes, that is correct, but they had to comply
with certain requirements that were laid out in the agreement,
and not all high-performance computers that were shipped
without a license to the civilian sector complied; so that was
also a problem.
Chairman Thompson. Have you seen this agreement?
Mr. Johnson. No.
Chairman Thompson. Who has seen this agreement?
Mr. Johnson. Well, it is a classified agreement, and we do
have access to classified information, of course; but I have
not personally read the agreement.
Chairman Thompson. Well, are you familiar with the Cox
Report's reference to this?
Mr. Johnson. Yes, I am.
Chairman Thompson. Is it not true that the Cox Report says
there is such an agreement; that the administration would not
release the agreement because the Chinese would not allow it;
and that the Cox Committee had reviewed the agreement and found
it to be wholly inadequate?
Mr. Johnson. I read that in the Cox Report, yes.
Chairman Thompson. Do you know of any national security
reason why the American people should not see the agreement
that allows us to do post-shipment verifications for high-
performance computers?
Mr. Johnson. Not to my knowledge.
Chairman Thompson. Does anybody else know who might have
access to this agreement or know anything about what is in the
agreement?
Mr. Johnson. No, sir.
Chairman Thompson. Well, I think that this is something
that we might want to inquire about, because a large part of
what we are doing is supposed to be in reliance upon the fact
that computers that are sent for civilian purposes, for
example, do not wind up in civilian hands. If the Cox Committee
has concluded that this agreement is wholly inadequate and the
administration will not release the agreement because of
Chinese objections, I think that raises serious concerns. I
think it might be a good idea perhaps to inquire of the
administration whether they would let us review that agreement
to see whether there are any legitimate national security
purpose for withholding portions of it, and that certainly can
be dealt with. But I see no reason on something like a post-
shipment verification arrangement and the extent to which we
should be able to rely upon who the real end-user is going to
be in any given situation, why this information should be
withheld from the American people.
Mr. Hoydysh, what is the economic effect of what we are
talking about? I have seen numbers that seem to indicate that,
right now anyway, there are only a limited number of high-
performance computers sold to Tier III countries. Obviously,
you have large domestic sales. You have large international
sales, most of which are not controlled because they are not of
a certain level.
Can you give us some feel in terms of numbers of sales or
potential sales that we are talking about--again, not that you
cannot export to Tier III countries, but that for some items
you have to go through a license process? I guess you have the
delay issue and then you have the denial issue, both; but can
you give us some better idea as to what the commercial impact
of this is for the computer industry, which clearly has
millions of sales domestically and internationally?
Mr. Hoydysh. Senator, in terms of the absolute volumes and
the absolute sales, this is not an overwhelming market for us
at this time. Depending on the company, depending on how you
slice it, it is anywhere from 10 to 15 percent of company
sales; that is all of Tier III. Now, China is the fastest-
growing market in Asia, and I believe is, after Japan, the
second-largest market. It certainly has tremendous potential.
We do not--again, just to clarify--we do not object to
licensing requirements for truly high-end systems. In fact, I
do not believe that any truly high-end system has ever been
approved to China. What we are concerned about is competition
at the lower end with commodity machines which are available
from a wide variety of sources, including from companies in
China.
The problem is one of delay, even without licenses, you
have a 10-day delay. But, on top of that, the United States is
the only country that requires our vendors to get an end-user
certificate from the Chinese government; so that a comparable
sale from anyone else would not require an end-user
certificate, and this is a bureaucratic process that can take
anywhere from 2 to 6 weeks.
Chairman Thompson. Have you been following any of the
hearings and debates that we have been having here on the PNTR
with regard to China, and national missile defense? And are you
familiar with the testimony that we have heard from this table,
from our intelligence officers, including the CIA, giving their
biennial estimates on weapons of mass destruction and nuclear
proliferation, in which they said that China still is the
greatest proliferator of weapons of mass destruction, and that
Beijing is continually supplying the rogue nations that are
increasingly becoming a threat to this country with regard to
biological, nuclear and chemical capabilities?
We hear that testimony all the time. That is why I consider
it more than a bureaucratic quibble to require some kind of
end-user satisfaction with regard to the Chinese. This is why
it concerns me that the administration wants to keep under
wraps the agreement that we supposedly have entered into as to
the way we are supposed to have some satisfaction on post-
shipment verification.
So it is not just strictly a Chinese deal. It is a
complicated world out there, and part of what is happening is
that the Chinese are supplying dual-use items, technicians,
technology, raw materials, and components, to a host of rogue
nations. Now, that is not your business, but that is our
business, and I just want you to know that we are not trying to
be unduly restrictive or anti-competitive, and we understand
the genie is out of the bottle. But this is part of what we
have got to balance, and this is why we have got to be careful
as, on the one hand, we try to embrace these countries, and
trade, and get along as best we can, and, on the other, we
remain mindful of their threats and of information revealed by
our own intelligence analyses, whether it be by the Rumsfeld
Commission or the Deutsch Commission or our own intelligence
assessments. These assessments are continually saying that this
country is doing things that pose a threat to our national
security, albeit, in some cases, in a roundabout way through
the rogue nations.
So that is a part of this process, too, and that has got to
enter into the balance. It looks to me like this has front-end
and back-end ramifications in terms of this process. It is the
process, it seems to me, that is most important here. No one
can sit here--I cannot sit here--and say what the MTOPS level
ought to be. But I do not think somebody within the bowels of
the Department of Commerce ought to be the unilateral
determiner of what that MTOPS level should be, either.
I think we need a process that includes all of the relevant
people at the table. I will make a final point here. It seems
to me that we should not overlook the fact that our allies, our
economic competitors and other countries in which these
computers are made and so forth, they have licensing processes
themselves. It's also worth remembering that when you are
looking at something like foreign availability or mass-
marketing, some are proposing taking whole categories of things
out of the process--decontrolling them because of alleged
foreign availability.
We are not even looking at the question of whether or not
``foreign available'' computers are available perhaps only to
our competitors' licensing process. If they have a licensing
process and they control these items, even if it is ``foreign-
available,'' do we want to totally decontrol this item? When
they have got a licensing regime and we will not have one
anymore, aren't they going to immediately do away with their
licensing regime too? It is going to be a race to the bottom.
So those are the sort of things that concern me. I am
making more of a statement. Anybody can comment on any of that,
if you want to.
Mr. Hoydysh. Let me just make a comment on the question of
foreign availability, and I think that has to be split into two
parts. There are two issues here; one, can other countries make
these products? And I think that, we can demonstrate very
clearly that almost anyone can make them. There is no
technological impediment to making them--it is a economic
impediment.
Everyone can make them, but not everyone can make money
making them; so that is what keeps people from getting into the
business. The other question is whether they have equal export
controls. My experience, and I have been doing this for 17
years, including attending COCOM for 7 years, that our allies
stayed in COCOM only by the force of the will of the United
States. If we had opened the door, they would have been out of
COCOM way before it was formally dissolved.
One of the big differences between COCOM, which at least
worked fairly well, and today is that there was then general
agreement about a common threat. Right now, there is little
agreement on the common threat. There is agreement on the rogue
nations, but there is no agreement that China poses a threat.
In fact, the other COCOM countries explicitly--the other
Wassenaar countries explicitly rejected putting China on a
target list.
Another complicating factor that one of our targets in Tier
III, Russia, is a member of Wassenaar. In addition, other
members of Wassenaar are actually trying to get China to join
Wassenaar.
So if you have China and Russia as members of Wassenaar,
which is the organization that is supposed to control exports
to these countries, you end up with some very strained
relationships and situations.
Chairman Thompson. Well, that is interesting you should
mention that, because I just came back a few weeks ago from a
trip. I went to Vienna, talked to the Wassenaar Arrangement
people there, talked to several of our allies about the
Wassenaar Arrangement and what they thought about it, and ran
into some of the things that you are talking about. There is
more than one view as to the COCOM situation and as to who the
leader was in disbanding that, however, and a lot of people
think it was the United States. Some people think it should
have been, and certainly something different should have come
about.
So we have got something that really is dysfunctional in
many respects, and that is the Wassenaar Arrangement. You are
right; Russia is a part of that. And some of our allies too:
Oftentimes, we have particular problems with the French,
amazingly enough, and we and they do not see eye-to-eye on many
things. The question is, what kind of example are we going to
set; and, what are we doing particularly with regard to these
high-performance computers?
I was surprised to hear from some of our allies while I was
over there that far from playing catch-up to our competitors,
the United States is leading the way in decontrolling these
high-performance computers, much to the chagrin of some of our
Japanese friends and our European allies. I was also surprised
to learn that our allies still, in some cases, maintain some
controls on these computers that we are not controlling; and
that we are constantly the ones that are pushing the limit
because of the competitive advantage we have there.
So I do not think that this is a totally black-and-white
picture. The other thing I would ask is whether or not there is
any validity to the notion that it makes a big difference
whether one secretly steals a computer or whether one buys it
legitimately. Sure, many countries have substantial
capabilities in terms of high-performance computers, whether
they are making them themselves or import into them, either
legally or illegally. But that is much different than having a
formal relationship with a legitimate supplier, in which you
are exporting mass quantities and where the recipient is able
to receive the technical support and training.
Does it not slow them down to be denied such support? I
think many times maybe this is what we are talking about, what
we are trying to do while we are building the national missile
defense system and reconfiguring our military and all that.
Perhaps we are just talking about slowing down the
proliferation process with regard to problem countries. Does it
not slow down the process if you deprive them of the technical
support and training?
We maintain our ability to track who gets what by having a
control system--track what is going out of the country; we do
not wholly decontrol; we do not give to just anybody the
technical support and training that would go with a traditional
commercial transaction. Is there any validity to that concept,
do you think?
Mr. Hoydysh. Absolutely, Senator, there is validity to it.
But just to put this in context, we are not asking to release
all high-performance computers. We are only asking for easing
of restrictions on the lowest level of high-performance
computers, assuming for the moment they can be still called
high-performance computers. These are common business servers
which are used in electronic commerce, and we have talked about
a level of four-to-eight processors. Each of our companies make
computers that have 32, 64, and above, and the highest
computers, the most powerful ones, have thousands of
processors.
So we still believe in strict controls on that level of
technology, and not even a question of licensing. We do not
even think that the highest end should be exported to Tier III,
period. What we are talking about is the large-volume, lower-
end, which is absolutely essential if China is going to develop
its Internet infrastructure, if it is going to develop its e-
commerce capabilities, which would allow it to have more
information, and more interaction. These computers provide the
backbone for that system infrastructure.
We are not talking about all computers. We are talking
about the lowest slice of technology that is widely available.
Chairman Thompson. And we are not talking about stopping
the sale of those computers. We are talking about a review
process.
Mr. Hoydysh. And that is where the difficulty lies, because
at this level of processor, you are talking more and more about
direct sales, about delivery in a matter of days. These are not
things that take a long time to build. Most of these things are
built within a few days of receiving the order; so speed of
delivery is essential.
One of the biggest problems with the review process that we
have is that even if it lasts 10 days, you have to add the
Chinese end-user certificate process to it. Then you have to
add the fact that each of these exports requires, by law
visitation by government official on top of that. Foreign
rendors can deliver the same product at relatively the same
price with the same performance without all of this
bureaucratic baggage.
It makes our stuff less competitive and, in addition to
which, we cannot utilize third-party distributors because we
cannot ship the product into China unless we have the end-user
identified at the time of shipment; so that we are deprived of
a whole channel of distribution of these systems.
Chairman Thompson. There are a lot of distribution problems
in selling goods to China that have nothing to do with our
controls on this end.
Mr. Hoydysh. No, but this is a specific aspect of the way
some of our companies do business. They identify someone who
will provide the service, who will hold the product and
distribute it to individual end-users. We cannot use that
channel.
Chairman Thompson. One more thing. Your criteria seems to
be the sale of processors. We have been talking about MTOPS
levels.
Mr. Hoydysh. What I wanted to do was de-mystify the
question of MTOPS, because everyone is fixated on the 2,000
MTOPS level which, in the mid-1900's or early 1990's,
represented a powerful machine. Today, an Intel personal
computer, the Apple personal computers, have MTOPS rating of
almost 2,500. In addition, as some of the other folks here have
testified, MTOPS may not be a valid measure of national
security concern. Just because----
Chairman Thompson. Excuse me. But for better or for worse,
the MTOPS level still is the criteria we are having to deal
with here. So what are you suggesting be decontrolled in terms
of MTOPS levels?
Mr. Hoydysh. At MTOPS levels now, coming out with the
newest processors that will be available later this summer----
Chairman Thompson. Again, we are looking into the future a
little bit.
Mr. Hoydysh. Well, you see, you are right, we are looking
into the future; but what makes it so much more difficult for
us is, because of the 6-month delay and because it takes at
least 3-plus months for an interagency process to come up with
a number, we are required to forecast what would be available
anywhere from 9 months to a year in advance.
In the last go-round, we actually missed the fact that the
Apple Macintosh arrived sooner than expected and could not be
sold through these distribution channels. So that one of the
reasons for shortening this 6-month congressional review period
is to make that forecast less prospective.
We are not saying decontrol today that which will be
available 6-months from now.
Chairman Thompson. I understand. I did not mean to get you
diverted. We were getting back to the MTOPS level. Can you
translate what you would suggest?
Mr. Hoydysh. We are suggesting that if the MTOPS level is
announced in July, let's say, and it becomes available 6 months
from now, which is still the current law, then we need to be
able to sell four-processor systems, which are large-volume
systems made with the new Itanium microprocessor, and that
number, and I could be off by several hundred, is somewhere on
the order of 27,000 MTOPS. That is a four-processor system.
Later on next year, and this is January 2001, mid-July,
2001, that number goes up to somewhere on the order of 33,000
or 34,000 MTOPS. And I do happen to have a prop----
Chairman Thompson. I thought as of February of this year,
you were at the 20,000 MTOPS level for civilian use anyway for
Tier III countries.
Mr. Hoydysh. Well, the level that we are talking about is
the lower level. That is the level below which government
review is not required.
Chairman Thompson. Right.
Mr. Hoydysh. So, right now, that level is 6,500 and the
civilian level is 20,000. But above 6,500 we have to submit it
to the government for a 10-day review, so that the government
still has the ability to review the end-user above 6,500 MTOPS.
Chairman Thompson. Maybe February was an announced date
instead of an effective date.
Mr. Hoydysh. It is an announced date. That does not----
Chairman Thompson. That was 12,500 for military.
Mr. Hoydysh. That becomes effective August.
Chairman Thompson. OK.
Mr. Hoydysh. Because of the 6-month delay.
Chairman Thompson. All right.
Mr. Hoydysh. Any announcement in July would become
effective in January 2001.
Chairman Thompson. Thank you very much. Senator Lieberman.
Senator Lieberman. Thanks, Mr. Chairman. I want to come
back to Mr. Johnson and Mr. Lieberman briefly. Mr. Johnson, you
said at one point in your testimony that post-shipment
verifications are important, but if I heard you correctly, they
do not always tell us what the end-use of the computer is.
Mr. Johnson. That is correct.
Senator Lieberman. So what I wanted to ask you was what
system would you put in place--or would you--to help us better
do verification after shipment to see exactly how these items
are being used?
Mr. Johnson. Well, I think our conclusion is that post-
shipment verification is an important process, because it does
identify the location of the computer. You can tell what kind
of facility it is in, but you cannot necessarily determine what
it is being used for. That requires some highly-trained
technicians to be able to go in and look at the data that is in
the computer, the computer codes, the programming and all, to
determine how that computer is being used.
We do not have a fix for that, but we do think the process
of just identifying having that verification that the computer
is there does at least help keep the system honest. There may
be some occasions when the Department of Commerce would want
to, and I think it has on some occasions, used highly-trained
technicians from the national labs to look at how computers are
used, but that would be one alternative.
It is a very expensive process. These people are highly
paid and it takes time to do that.
Senator Lieberman. Is it clear that we have the authority
under law to do that next level of post-shipment verification
to see exactly how the computers are being used?
Mr. Johnson. I would have to research that. I think that we
do have that authority, but whether or not we would be able to
get the cooperation of the----
Senator Lieberman. The purchaser?
Mr. Johnson [continuing]. Government, the purchaser, to do
that, is another thing.
Senator Lieberman. So, in a way, you are saying that the
difference here is between determining where the computer ends
up and how it is used.
Mr. Johnson. How it is used; yes.
Senator Lieberman. OK. I would welcome any response you
have to that in writing afterwards.\1\
---------------------------------------------------------------------------
\1\ The information referred to appears in the Appendix on page
119.
---------------------------------------------------------------------------
Mr. Johnson. Sure.
Senator Lieberman. Which is whether the authority is there.
Mr. Lieberman--every time I say that, I feel as if I am having
a conversation with myself, which are some very good
conversations I have, of course, regularly. You talked about
the fact that you do not want to see any items dropped from the
control list without DOD approval at one point in your
testimony.
Just help me remember to what extent DOD participates with
Commerce in the construction of the control list; that is, the
dual-use items that are on the control list?
Mr. Lieberman. Well, currently, DOD basically generates a
national security list that is a list of militarily-critical
items, and then gives it to Commerce to be incorporated into
the Commerce control list, and we think that process works
pretty well. So, basically, what we are suggesting is simply
that this process be retained, or at least the essence of the
process be retained.
Senator Lieberman. You are really right at the heart of
this dilemma that we have talked about all morning, which is
here are these extraordinarily capable computers and other
items, and how do we determine how they are going to be used
and whether, in fact, they are--so how does the Department of
Defense make that judgment? Is it a cautious judgment? In other
words, is it a sort of worst-case scenario judgment, that here
is something that of course can be used for peaceful,
commercial purposes, but, in these circumstances, it is
possible that it could be used in a way that would threaten us?
Mr. Lieberman. Well, I would hesitate to make a general
characterization like that. Certainly Defense is subject to the
same pressures in terms of different opinions, different
inputs, from across the spectrum. Industry certainly provides
input to the department, and right now, a lot of the
department's efforts to re-engineer its own internal review
processes for export controls are largely driven by complaints
from industry that our process is inefficient and takes too
long.
So we are aware of that end of the spectrum. Of course, we
have several Defense agencies involved. The intelligence
community certainly inputs. But we are talking about a dynamic
situation where what it makes sense to control today may not
make a whole lot of sense down the road; and, in fact, the
control list does change over time.
I think there is certainly a legitimate case to be made
that the control list ought to be under constant scrutiny and
evaluation from the standpoint of advancing technology.
Senator Lieberman. And maybe that is not happening
frequently enough now or regularly enough now.
Mr. Lieberman. Well, I do not have enough knowledge of that
to say whether it is happening--we really have not made any
attempt to look into individual determinations of what has gone
on the list and what has come off the list. In fact, that is
the primary subject of next year's interagency IG review. You
may recall that the authorization act last year requires the
IGs of several Federal agencies to look at this whole process
annually for 7 years; and this year, we looked at what are
called deemed exports. Next year, we are going to be looking at
the composition of the control list, and I hope that I can give
you a much better answer perhaps this time next year.
Chairman Thompson. Excuse me. And also, would that include
who decides what goes on and comes off of it?
Mr. Lieberman. Yes, sir.
Chairman Thompson. OK.
Senator Lieberman. Thank you. We will look forward to that.
Mr. Milhollin, I want to come back to something you said,
which is if the current definition of foreign availability--I
presume, in 1712, continues--that we will, in fact, be
required--the United States will be required--to decontrol
certain items that our allies now control. Could you just
develop that thought a little bit more?
Mr. Milhollin. Yes. Well, to begin with, there is a
generally agreed list of things which we control in common with
our allies for each kind of technology.
Senator Lieberman. Right.
Mr. Milhollin. I am sure the Committee is familiar with
those. The five items in my testimony that I selected as
examples are controlled by our allies, as well as ourselves. I
think some of them are probably controlled for missile, as well
as nuclear, reasons. In my judgment, I think that it would be
very likely that foreign availability determinations would be
made for all of them, because they are made by manufacturers in
more than one country.
Senator Lieberman. Right. These are all again controlled
now by our major allies and ourselves?
Mr. Milhollin. Yes, and I think you could probably find
cases where foreign countries or countries that we are worried
about, say controlled countries, had managed to buy these
things on the world market. So, again, if you come down to the
position that if a controlled country can buy these things from
somebody, then they should not be controlled here, you run into
the problem that you would have to go down the whole control
list to see which items are available to a rogue from some
rogue supplier--I am sorry--available to a controlled country
from a rogue supplier, then you would have to make a judgment
in each case to what extent it is available, and our experience
shows that the controlled countries can get some of these
things some of the time from rogue suppliers.
For example, the Pakistanis have been quite successful in
importing missile technology of all kinds from China, and Iran
has been successful in importing poison gas technology from
China. In fact, lots of countries have been successful in
importing lots of things from China; and if you use that as a
standard, then you are going to have to decontrol a fair number
of items that our allies and we now control.
Senator Lieberman. Do you have a recommendation for a
better definition of foreign availability?
Mr. Milhollin. Well, we have a foreign availability
procedure now that has been criticized. I question whether--it
seems to me that if an exporter can go through the present
process and prove that something is foreign-available, then he
is entitled--then the exporter is entitled to some
consideration.
I have not sat down and tried to draft standards of my own.
I mean, it took me a fair amount of time to go through the
standards that are in the bill and compare these items to that
standard.
Senator Lieberman. Sure.
Mr. Milhollin. But I could say one thing, that the standard
that the Banking Committee has adopted seems to me to be
entirely too broad and too sweeping. Senator, if I could, I
would like to comment on something.
Senator Lieberman. Before you go there, let me just say
that if you have the time and inclination, I think it would be
very helpful if you had some suggestions about what a better
standard might be than the one that is in the Banking Committee
bill.
Mr. Milhollin. Very well.
Senator Lieberman. Please go ahead with what you were going
to say.
Mr. Milhollin. I was going to say a couple of things in
response to Mr. Hoydysh's answers to the Chairman's questions.
The most recent data on the amount of supercomputer exports to
Tier III countries is about 5 percent.
Senator Lieberman. Five percent of?
Mr. Milhollin. Five percent of the supercomputer market or
the high-performance computer market.
Senator Lieberman. And that is the world market or the
American share of the business?
Mr. Milhollin. The American, I guess--5 percent of what we,
the United States, export.
Senator Lieberman. OK.
Mr. Milhollin. Because there are some numbers on that, and
the most recent ones I have seen put the share at about 5
percent. Second, I think we have talked a fair amount about
delays. The Commerce Department is now meeting its time
requirements in over 90 percent of its cases. So the Commerce
Department now has a pretty good record of getting dual-use
items turned around in a pretty good period of time.
The primary reason for that is that we are only controlling
about a tenth as much dual-use equipment as we controlled
during the Cold War; that is, in about 1989, we were
controlling about ten times as much as we are doing now. So,
with the lighter licensing burden, Commerce is able to turn
around the applications within its time restraints in about 90
percent--over 90 percent of its cases.
The third thing I would like to point out is we are not
talking about barring exports; we are talking about licensing
them.
Senator Lieberman. Right.
Mr. Milhollin. And, again, over 90 percent of the
applications are approved. So, for Mr. Hoydysh's purposes, I
would suggest to him that it is a good thing, if you are
exporting a sensitive item, to get the government to tell you
whether it might be going to the wrong place; that is, if I
were an exporter and I had the government giving me a free
bureaucrat that would tell me within 10 days whether my
customer was a problem, I think I would want to take the
government up on that, rather than read in The New York Times
or The Washington Post that my product had gone astray.
I think we are providing a good service; that is, a 10-day
review to tell an exporter, ``Look, you know, there is a
problem with this guy,'' or there is not. It seems to me an
exporter would--I do not understand why exporters do not want
that service. Let's put it that way.
Senator Lieberman. Mr. Milhollin, I want to finally just
ask you to respond to two other points made here, in some ways,
both by Mr. Hoydysh, but the first one is a general point that
is certainly made by those who support the current movement of
our export control system, which is to turn the national
security argument around, if you will, and say that at the
heart of our national security today is our technological
capability.
Part of the way the high-tech industries in America stay
strong is by enjoying a good share of the global market, and
if, in some sense, export controls are applied so rigidly or
demandingly that we deny them that market, that the effect will
be that they will have less resources with which to develop the
capabilities that make us a strong Nation. So how do you
respond to that?
Mr. Milhollin. I think that argument would be a valid
argument if the countries we are worried about were a major
part of their market. But, in fact, they are not.
Senator Lieberman. Including China?
Mr. Milhollin. Including China. Again, the most recent
figures I have seen show that Tier III, that is, the countries
we are worried most about for supercomputer exports, are taking
about 5 percent of our sales.
Senator Lieberman. Tier IV are what we more typically call
the rogue nations, Iraq, Iran, and Libya?
Mr. Milhollin. That is right.
Senator Lieberman. And then Tier III is China, Pakistan----
Mr. Milhollin. India, Israel, and Russia, that sort of
thing.
Senator Lieberman. Right.
Mr. Milhollin. I believe that in the supercomputer
industry, the companies--there are not very many companies.
There are six, eight, or ten. They are going to survive or not
depending on how they do in the big market; that is, the U.S.
market, the Japanese market, the European market, the markets
for truly civilian applications of high-speed computing.
They are going to make it or not with respect to each other
depending on how they do in those markets, not whether they
make a marginal sale to Tier III or not. So I think that it is
a good argument that we have to be strong and maintain our
competitive edge, but it is just that the numbers are not
there. Who makes it or does not is not going to depend on sales
to Tier III. That is my response.
Senator Lieberman. The second one was the very interesting
exchange between Senator Thompson and Mr. Hoydysh, and it goes
back in a way to something you said earlier in the initial
argument you made about America setting the standard, which is
that we did during the Cold War. That is part of why COCOM
worked and why, though there was naturally some leakage,
nonetheless, the former Soviet Union was impeded in its
development of some sophisticated systems.
And, of course, the argument would be as it was made by Mr.
Hoydysh, which is that the world has changed and we are post-
Cold War. Not only is it not a bipolar world anymore, but more
to the point here, though we have a rough consensus with our
allies about the rogue nations, the Tier IV nations--and,
again, there is some leakage there about Iran, Iraq, and Libya,
from some of our allies, in Europe particularly--the real
controversy seems to be over China and our differing attitudes,
notwithstanding some of the testimony today, from you
particularly, about China's proliferation activities.
So I wanted to give you a chance to update your argument
about the effectiveness of COCOM because of our American
leadership in a world that is quite different from the one in
which COCOM existed and, most particularly, in which we seem to
have some fundamental disagreements with our allies,
sophisticated, well-developed allies, about China.
Mr. Milhollin. I think that Mr. Hoydysh's point is a good
one. He argues that we do live in a different world and it is
true, there is less consensus and the targets of our activities
are not as well-defined, and, in particular on China, there is
a debate. But, you know, there is also a debate on Iran. I have
talked to German export control officials high up who do not
see Iran as a threat. In fact, one of them told me that Iran
was his favorite country.
We are in a world where one country's rogue is another
country's good customer.
Senator Lieberman. Right.
Mr. Milhollin. So this whole effort is much more difficult
and it is going to require much more aggressive and more
effective diplomacy by us than in the old days when it was
easier. But if you look at the alternatives, do we have an
alternative to doing it? I do not think we do. I think we have
to do the best we can in a new world which is more difficult,
but I do not think we can just say, ``Well, gee, the world is
really difficult now. It is very dangerous. We are just going
to throw up our hands and everybody is going to sell everything
to everybody and we are going to have total democracy in all
the technologies that are necessary to build weapons of mass
destruction.'' I fear that that is the tendency we are seeing,
but I do not think we are ready to live in the world of 1914 in
which everybody has the bomb.
Nuclear weapons grew up during the Cold War, which was a
pretty stable period, looking back on it now. If you postulate
the kind of--lots of countries with lots of different alliances
that we had before World War II, and you imagine lots of those
folks with nuclear weapons, we are not ready for that, but I
think that is where we are going. And what I am trying to argue
is that we should slow it down as much as we can.
Senator Lieberman. Sure. Understood. Mr. Hoydysh, do you
want a word to respond?
Mr. Hoydysh. Yes, very briefly. I am not going to argue
about whether it is 5 or 6 percent of the market. If we
believed that what we were doing was hurting national security,
it would not matter whether it was 5, 10, or 15 percent of the
market. We are convinced, though, that even the 5 percent,
which is bound to grow--Tier III countries represent about half
of the population of the world--we cannot afford to give up
those markets without having some serious impacts on our
technological leadership and on the health of our industry.
Five percent sounds like a small number, but what if someone
proposed to cut the defense budget by 5 percent? That has
significant impacts on our R&D and significant impact on where
we can compete.
We are proposing what we are proposing because we think on
balance it helps us more than it helps any potential enemies,
and that, even if we did not sell a single one of these items
that we are talking about--and I am not talking about high-end
computers, only the ones that we are talking about
decontrolling--that the target countries could get as many of
these as they wanted from other sources and we would have
accomplished nothing, other than losing 5 percent of the
market.
Senator Lieberman. That frames the issue and the difficulty
of our decisions. Thanks very much to all of you and to you,
Mr. Chairman. I think it has been for me a very helpful
morning.
Chairman Thompson. Well, thank you. Listening to you and
Mr. Milhollin, it seems to me what happened was that during the
Cold War, we had this pretty tough regime, this COCOM regime.
Then, the Cold War was over. We disbanded COCOM and we had a
lull period there. Now what has happened is that a new, more
diverse threat has emerged, in terms of the rogue nations. And
all we are left with is Wassenaar, which is very, very weak,
and we are struggling to see how much further we want to or can
go in terms of something less than COCOM, but more then
Wassenaar. It seems to me that this is what we are struggling
with.
There is one more point I would like to make before
concluding, one that I think is a very, very important one. It
concerns the idea of reducing the congressional renew period
for computer decontrols to 30 days, that is 30 calendar days
that is being proposed. If we were out of session, there would
be no review time at all. GAO would not even get it,
presumably. So that is somewhat unusual and, I think,
absolutely undesirable.
But finally, on a note of harmony, we had IG reports last
year, and we had an array of all the inspector generals before
us who looked at our export situation with regard to various
departments. There are clearly some things that we ought to be
doing that we are not doing, and that are not or should not be
controversial. We do not have enough licensing officers. They
apparently are not sufficiently trained. The law requires
training programs for these agencies. It is not being complied
with. The law requires a cumulative effect analysis that is not
being done. Nobody knows what the cumulative effect of all this
is. We look at these things one at a time.
We have in our export control bureaucracy, as we do in most
all the other government agencies, totally inadequate
information systems. Our computers do not talk to each other
with regard to this licensing process. Our law enforcement
people, who might have information on some of these entities
that our exporters are trying to deal with, and not integrated
sufficiently into the process. It is either not there, or not
used, and there is no coordination. It's a real management
problem.
That is what we ought to be doing first, I think. There is
an awful lot of stuff that we could do that would speed up the
process and also improve the safety of the process and help fix
some of the things that we are concerned about. So that should
be on the table, also. So, with that, we will cease and desist.
Thank you very, very much for this very enlightening hearing
that we have had today and your testimony.
The record will remain open for a week after the close of
the hearing. So, we are in recess.
[Whereupon, at 12:20 p.m., the Committee was adjourned.]
A P P E N D I X
----------
PREPARED STATEMENT OF SENATOR AKAKA
Mr. Chairman and Senator Lieberman thank you for holding this
hearing.
How to control exports critical to our national security in a world
of rapid technological innovation is one of the most serious issues we
face.
I hope that this Committee will hold more hearings on this topic
and I would recommend looking at the administration's new proposals on
export controls announced just this week.
Most of us would consider computers to be on the cutting edge of
technologies which we should control. But at the same time the
definition of ``cutting edge'' is constantly changing. Sometime this
year Intel will introduce a new chip which will more than double the
current level of computer processing capability.
Efforts to control this technology sometimes become ridiculous. For
example, this fall Sony will introduce its new PLAY STATION II which
contains a processor above the performance levels set by current
Japanese export controls. Rather than restrict PLAY STATION exports,
the Japanese redefined how to control such items.
We are in a similar situation in this country. Every few years--
with increasing frequency--every administration since President
Reagan's has had to revise controls on computer exports.
This has become an even more critical question as the American
computer industry earns more than 50 percent of its revenues from
exports. With the speed of innovation and the need to protect market
share from foreign competition, I can see why the industry is eager to
raise the level of permissible exports and speed up the license review
process.
This is an industry in which innovation is the key to market
success. American manufacturers do not have a monopoly on production.
For example, 80 percent of all computer motherboards are manufactured
in Taiwan. One of the fastest growing computer companies in the world
is in Beijing. To keep pace with this competition, American
manufacturers need the revenues to plow back into research and
development.
In December 1999, the Pentagon's Defense Science Board Task Force
on Globalization and Security concluded that ``if U.S. high-tech
exports are restricted in any significant manner, it could well have a
stifling effect on the U.S. military's rate of technological
advancement.''
In effect, this is the heart of the problem: How do we control
critical defense exports without stifling the innovation necessary to
national security in a world in which the globalization of technology
can outstrip our ability to control it?
I look forward to the witnesses today and their answer to this
question.
[GRAPHIC] [TIFF OMITTED] T5171.001
[GRAPHIC] [TIFF OMITTED] T5171.002
[GRAPHIC] [TIFF OMITTED] T5171.003
[GRAPHIC] [TIFF OMITTED] T5171.004
[GRAPHIC] [TIFF OMITTED] T5171.005
[GRAPHIC] [TIFF OMITTED] T5171.006
[GRAPHIC] [TIFF OMITTED] T5171.007
[GRAPHIC] [TIFF OMITTED] T5171.008
[GRAPHIC] [TIFF OMITTED] T5171.009
[GRAPHIC] [TIFF OMITTED] T5171.010
[GRAPHIC] [TIFF OMITTED] T5171.011
[GRAPHIC] [TIFF OMITTED] T5171.012
[GRAPHIC] [TIFF OMITTED] T5171.013
[GRAPHIC] [TIFF OMITTED] T5171.014
[GRAPHIC] [TIFF OMITTED] T5171.015
[GRAPHIC] [TIFF OMITTED] T5171.016
[GRAPHIC] [TIFF OMITTED] T5171.017
[GRAPHIC] [TIFF OMITTED] T5171.018
[GRAPHIC] [TIFF OMITTED] T5171.019
[GRAPHIC] [TIFF OMITTED] T5171.020
[GRAPHIC] [TIFF OMITTED] T5171.021
[GRAPHIC] [TIFF OMITTED] T5171.022
[GRAPHIC] [TIFF OMITTED] T5171.023
[GRAPHIC] [TIFF OMITTED] T5171.024
[GRAPHIC] [TIFF OMITTED] T5171.025
[GRAPHIC] [TIFF OMITTED] T5171.026
[GRAPHIC] [TIFF OMITTED] T5171.027
[GRAPHIC] [TIFF OMITTED] T5171.028
[GRAPHIC] [TIFF OMITTED] T5171.029
[GRAPHIC] [TIFF OMITTED] T5171.030
[GRAPHIC] [TIFF OMITTED] T5171.031
[GRAPHIC] [TIFF OMITTED] T5171.032
[GRAPHIC] [TIFF OMITTED] T5171.033
[GRAPHIC] [TIFF OMITTED] T5171.034
[GRAPHIC] [TIFF OMITTED] T5171.035
[GRAPHIC] [TIFF OMITTED] T5171.036
[GRAPHIC] [TIFF OMITTED] T5171.037
[GRAPHIC] [TIFF OMITTED] T5171.038
[GRAPHIC] [TIFF OMITTED] T5171.039
[GRAPHIC] [TIFF OMITTED] T5171.040
[GRAPHIC] [TIFF OMITTED] T5171.041
[GRAPHIC] [TIFF OMITTED] T5171.042
[GRAPHIC] [TIFF OMITTED] T5171.043
[GRAPHIC] [TIFF OMITTED] T5171.044
[GRAPHIC] [TIFF OMITTED] T5171.045
[GRAPHIC] [TIFF OMITTED] T5171.046
[GRAPHIC] [TIFF OMITTED] T5171.047
[GRAPHIC] [TIFF OMITTED] T5171.048
[GRAPHIC] [TIFF OMITTED] T5171.049
[GRAPHIC] [TIFF OMITTED] T5171.050
[GRAPHIC] [TIFF OMITTED] T5171.051
[GRAPHIC] [TIFF OMITTED] T5171.052
[GRAPHIC] [TIFF OMITTED] T5171.053
[GRAPHIC] [TIFF OMITTED] T5171.054
[GRAPHIC] [TIFF OMITTED] T5171.055
[GRAPHIC] [TIFF OMITTED] T5171.056
[GRAPHIC] [TIFF OMITTED] T5171.057
[GRAPHIC] [TIFF OMITTED] T5171.058
[GRAPHIC] [TIFF OMITTED] T5171.059
[GRAPHIC] [TIFF OMITTED] T5171.060
[GRAPHIC] [TIFF OMITTED] T5171.061
[GRAPHIC] [TIFF OMITTED] T5171.062
[GRAPHIC] [TIFF OMITTED] T5171.063
[GRAPHIC] [TIFF OMITTED] T5171.064
[GRAPHIC] [TIFF OMITTED] T5171.065
[GRAPHIC] [TIFF OMITTED] T5171.066
[GRAPHIC] [TIFF OMITTED] T5171.067
[GRAPHIC] [TIFF OMITTED] T5171.068
[GRAPHIC] [TIFF OMITTED] T5171.069
[GRAPHIC] [TIFF OMITTED] T5171.070
[GRAPHIC] [TIFF OMITTED] T5171.071
[GRAPHIC] [TIFF OMITTED] T5171.072
[GRAPHIC] [TIFF OMITTED] T5171.073
[GRAPHIC] [TIFF OMITTED] T5171.074
[GRAPHIC] [TIFF OMITTED] T5171.075
[GRAPHIC] [TIFF OMITTED] T5171.076
[GRAPHIC] [TIFF OMITTED] T5171.077
[GRAPHIC] [TIFF OMITTED] T5171.078
[GRAPHIC] [TIFF OMITTED] T5171.079
[GRAPHIC] [TIFF OMITTED] T5171.080
[GRAPHIC] [TIFF OMITTED] T5171.081
[GRAPHIC] [TIFF OMITTED] T5171.082
[GRAPHIC] [TIFF OMITTED] T5171.083
[GRAPHIC] [TIFF OMITTED] T5171.084
[GRAPHIC] [TIFF OMITTED] T5171.085
-