S. Hrg. 105-238
PROLIFERATION AND U.S. EXPORT CONTROLS
=======================================================================
HEARING
before the
SUBCOMMITTEE ON INTERNATIONAL SECURITY,
PROLIFERATION, AND FEDERAL SERVICES
of the
COMMITTEE ON
GOVERNMENTAL AFFAIRS
UNITED STATES SENATE
ONE HUNDRED FIFTH CONGRESS
FIRST SESSION
__________
JUNE 11, 1997
__________
Printed for the use of the Committee on Governmental Affairs
U.S. GOVERNMENT PRINTING OFFICE
41-566cc WASHINGTON : 1997
_______________________________________________________________________
For sale by the Superintendent of Documents, Congressional Sales Office
U.S. Government Printing Office, Washington, DC 20402
COMMITTEE ON GOVERNMENTAL AFFAIRS
FRED THOMPSON, Tennessee, Chairman
SUSAN M. COLLINS, Maine JOHN GLENN, Ohio
SAM BROWNBACK, Kansas CARL LEVIN, Michigan
PETE V. DOMENICI, New Mexico JOSEPH I. LIEBERMAN, Connecticut
THAD COCHRAN, Mississippi DANIEL K. AKAKA, Hawaii
DON NICKLES, Oklahoma RICHARD J. DURBIN, Illinois
ARLEN SPECTER, Pennsylvania ROBERT G. TORRICELLI, New Jersey
BOB SMITH, New Hampshire MAX CLELAND, Georgia
ROBERT F. BENNETT, Utah
Hannah S. Sistare, Staff Director and Counsel
Leonard Weiss, Minority Staff Director
Michal Sue Prosser, Chief Clerk
------
SUBCOMMITTEE ON INTERNATIONAL SECURITY, PROLIFERATION AND FEDERAL
SERVICES
THAD COCHRAN, Mississippi, Chairman
SUSAN M. COLLINS, Maine CARL LEVIN, Michigan
PETE V. DOMENICI, New Mexico DANIEL K. AKAKA, Hawaii
DON NICKLES, Oklahoma RICHARD J. DURBIN, Illinois
ARLEN SPECTER, Pennsylvania ROBERT G. TORRICELLI, New Jersey
BOB SMITH, New Hampshire MAX CLELAND, Georgia
Mitchel B. Kugler, Staff Director
Linda Gustitus, Minority Staff Director
Julie Sander, Chief Clerk
C O N T E N T S
------
Opening statements:
Page
Senator Cochran.............................................. 1
WITNESSES
Wednesday, June 11, 1997
William A. Reinsch, Under Secretary for Export Administration,
Department of Commerce......................................... 3
Mitchel B. Wallerstein, Deputy Assistant Secretary for
Counterproliferation, Department of Defense.................... 12
Stephen D. Bryen, President, Delta Tech.......................... 36
William Schneider, Fellow, Hudson Institute...................... 43
Alphabetical List of Witnesses
Bryen, Stephen D.:
Testimony.................................................... 36
Prepared statement........................................... 40
Reinsch, Willliam:
Testimony.................................................... 3
Prepared statement........................................... 8
Schneider, William, Jr.:
Testimony.................................................... 43
Prepared statement........................................... 45
Wallerstein, Mitchel B.:
Testimony.................................................... 12
Prepared statement........................................... 14
PROLIFERATION AND U.S. EXPORT CONTROLS
----------
WEDNESDAY, JUNE 11, 1997
U.S. Senate,
Subcommittee of International Security,
Proliferation, and Federal Services
of the Committee on Governmental Affairs,
Washington, DC.
The Subcommittee met, pursuant to notice, at 9:40 a.m., in
room SD-342 Dirksen Senate Office Building, Hon. Thad Cochran,
Chairman of the Subcommittee, presiding.
Present: Senators Cochran and Durbin.
OPENING STATEMENT OF CHAIRMAN COCHRAN
Senator Cochran. The Subcommittee will please come to
order.
Let me first welcome everyone to today's hearing of the
Subcommittee on International Security, Proliferation, and
Federal Services of the Governmental Affairs Committee.
The topic of today's hearing is Proliferation and U.S.
Export Controls.
The Subcommittee has previously held hearings on Chinese
and Russian Proliferation. Today we will examine how U.S. dual-
use export control policies may be promoting military
modernization in other nations, principally Russia and China,
and the extent to which this modernization helps these nations
continue their proliferant activities.
With the end of the Cold War came the need to reexamine
American export control practices, especially with respect to
goods having both military and civilian applications, goods
commonly referred to as dual use. During the latter stages of
the Cold War, approximately $100 billion per year worth of
exports required an export license.
In 1996, the Commerce Department licensed for export $4.9
billion worth of dual-use technology, while our total export
volume of goods and services amounted to $846 billion. The
licensed exports comprised just under six-tenths of one percent
of total U.S. exports in 1996.
Or, if you wish, look at American export controls in
another way. Less than \1/2\ of 1 percent of the U.S. economy
is covered by export controls. Of that, more than 95 percent of
export license requests are approved.
President Clinton entered office promising to liberalize
U.S. export controls. He restated this promise in a September
1993 letter to Edward McCracken, Chairman and CEO of Silicon
Graphics, when he said that he is ``personally committed to
developing a more intelligent export control policy, one that
prevents dangerous technologies from falling into the wrong
hands without unfairly burdening American commerce.''
I am concerned that President Clinton's relaxed export
control policy on supercomputers has accomplished precisely the
opposite of his stated intention of ``preventing dangerous
technologies from falling into the wrong hands.''
According to Victor Mikhailov, Russia's Minister of Atomic
Energy, we now know that five American supercomputers--four of
which came from Edward McCracken's Silicon Graphics--reside in
Russia's premier nuclear weapons design labs. We also know,
from Secretary Reinsch's testimony to the House of
Representatives in April, that 46 American supercomputers are
in the People's Republic of China, at least one of which was
sold to the Chinese Academy of Sciences by Silicon Graphics.
And these 46 may be only the tip of the iceberg.
We know some other things, too. We know, of course, that
Russia's nuclear weapons labs design nuclear weapons. We also
know that the Chinese Academy of Sciences is a key participant
in Chinese military research and development, and has been for
a long time, working on everything from the DF-5 ICBM--which is
capable of reaching the United States--to uranium enrichment
for nuclear weapons.
According to the Chinese Academy of Sciences, its new
Silicon Graphics ``Power Challenge XL'' supercomputer provides
the Academy with ``computational power previously unknown''
which is available to ``all the major scientific and
technological institutes across China.''
The good news is that some of these ``major scientific and
technological institutes across China'' may not be involved in
developing weapons of mass destruction and missile delivery
systems for China and its clients, but some surely are, and
they're doing this work courtesy of what appears to be, at
best, a deeply flawed U.S. export control policy.
Today the Subcommittee will examine what the dual-use
technology export control policies of the United States are,
how they are working, and to what extent these policies need
revision.
Our witnesses today are well-qualified to assist in the
examination of these questions. We will hear first from William
Reinsch, Under Secretary of Commerce for Export Administration,
who worked previously on Capitol Hill and, today, oversees
export control policy for the Commerce Department, and Dr.
Mitchel Wallerstein, who as the Deputy Assistant Secretary of
Defense for Counterproliferation, is the Defense Department's
most senior official with responsibility for these issues.
Our second panel is comprised of Dr. Stephen Bryen,
President of Delta Tech and a former Defense Department
official with the responsibility for export controls, and Dr.
William Schneider, a former State Department official who is
currently a Hudson Institute Fellow and also the chairman of
the State Department's Defense Trade Advisory Group.
The Cold War's end does not decrease the need for the
continued safeguarding of sensitive American dual-use
technology. While there may no longer be a single, overarching
enemy of the United States, there is little doubt that many
rogue States, and perhaps others, have interests clearly
inimical to those of the United States. Helping these nations--
or helping other nations to help these nations--to acquire
sensitive dual-use technology capable of threatening American
lives and interests makes no sense.
I am going to ask Secretary Reinsch to proceed with
whatever comments or statements he would like to make. We have
a copy of a statement, which the Committee has received, which
we will print in the record in full, and then we will hear from
Dr. Mitchel Wallerstein, and then have an opportunity to
discuss your testimony.
Secretary Reinsch, you may proceed.
TESTIMONY OF WILLIAM A. REINSCH, UNDER SECRETARY FOR EXPORT
ADMINISTRATION, DEPARTMENT OF COMMERCE
Mr. Reinsch. Thank you very much, Mr. Chairman. It is a
pleasure to be back in this room. On the whole, I think I would
rather be back up there sitting behind you, but that is all
right. I am looking forward to the dialogue.
The President considers an effective strategic trade export
control program to be a critical element of our overall
national security posture and, as you noted, he has directed us
to regularly update our system so that it focuses on the new
threats we face today.
Since the end of the Cold War, crafting export control
policy has become more difficult because the world is more
complex and the battle lines between competing interests less
defined. The Cold War had a certain elegant simplicity. The
United States and its allies had a clear enemy, and we largely
agreed on how it should be contained. Economic sacrifice was
often asked and usually made by countries and companies in the
name of containment, and that was a policy that worked.
That structure, though, has now been replaced by less-
defined and more ambiguous threats no longer confined to a
handful of relatively predictable actors. The immediate threats
are now terrorism and the proliferation of weapons of mass
destruction to a handful of smaller, geographically diverse
rogue States.
At the same time, the rapid spread of advanced technology
in a globalizing economy has made critical items widely
available--including some of the ones that we will be
discussing today--and it has greatly increased the number of
nations capable of producing advanced technology.
As a result, the United States does not have a monopoly on
these items, if it ever did, and it has become harder to reach
international consensus on what threats we face and harder to
enforce any agreements that we do reach. For many nations,
economic objectives are now paramount as they seek to penetrate
new markets. Yesterday's adversaries are today's customers, and
yesterday's allies are today's competitors.
Even when a policy is clear, our ability to implement it is
not. The world abhors chemical and biological weapons, for
example, but they can be produced with 40-year-old technologies
using feedstocks and equipment found in hotel kitchens,
breweries, universities and even high schools all over the
world.
Building a nuclear weapon, as you know, Mr. Chairman, does
not require sophisticated computers.
The administration's response to these changed
circumstances is based on four major principles.
The first is reforming the export licensing process so that
all relevant agencies can bring their expertise to the table in
a timely manner. This allows for comprehensive interagency
review of sensitive transactions while ensuring that the
process does not put U.S. exporters at a disadvantage.
Second is streamlining controls so they focus on items that
pose the greatest threat to our security;
Third is clarifying our regulatory regime so that exporters
know what their obligations are and know how to improve their
internal compliance programs; and
Fourth is strengthening the multilateral systems.
With regard to process reform--and I think this is
important in light of the discussion that we'll have today--we
have revamped the licensing process, via Executive Order, so
that all relevant agencies can review all export license
applications if they wish. In return for that expansion of
review authority, the other agencies have committed to Commerce
to conduct their reviews within strict time limits, to provide
a statutory or regulatory basis for their views, and to
participate in a dispute settlement process at appropriate
political levels.
It is important to note in that process that some 96
percent of the applications that we review are resolved by
interagency consensus at the working level. Those were there
are differences of opinion are by far the minority and they are
worked out in the dispute settlement process that I referred to
that is implemented in the President's Executive Order of
December 1995. Thus far all specific license disputes have been
settled and have not had to be escalated beyond the Assistant
Secretary level.
With respect to streamlining, we have updated controls on a
wide variety of equipment, including high-performance
computers, in order to reflect rapid technological advances
that have made previously controlled items old technology
widely available from numerous foreign sources.
Let me say a couple words about computers, if I may, Mr.
Chairman, in view of your remarks.
In 1992, we treated a computer capable of running at 195
Million Theoretical Operations per Second or MTOPS as a
supercomputer subject to strict controls. Today, personal
computers that exceed this level of performance are being sold
for less than $2,000 at retail stores such as Best Buy, and
Radio Shack, and through mail order catalogues.
Not having seen your back room in a while, Mr. Chairman, I
suspect the computers that are in there are supercomputers,
according to the 1995 definition, and I suspect the ones in
your personal office are in that same category.
When President Clinton took office, he was urged by
Congressional leaders of both parties to make long overdue
reforms in this area, and I believe our efforts to do that have
been a model of good government decision making. The
President's 1995 decision was the result of a joint interagency
recommendation based on work that various agencies, including
the Defense Department, did internally as well as a private
sector study. The studies came to similar conclusions--that
advances in computing technology were making ever-higher
performing computers widely available internationally to the
point where controls on them would be ineffective. In addition,
they concluded that the level of computer power needed for a
number of activities, including nuclear weapons development,
was already widely available abroad. Other functions, which we
wanted to protect, required performance levels well above the
levels that the President ultimately set in his decision of
October of 1995.
It is also worth noting that none of these studies took
full account of the rapid development of semiconductor
technology that has permitted significant upgrading of existing
machines by adding processors, as well as taking into account
parallel processing, the linking together of many smaller
computers to achieve the same effect as a much larger machine.
Both of these developments have had an enormous impact on
making high-performance computers essentially commodity
products. In 1996, for example, the average--not the highest,
but the average performance level for a multiple processor was
6,923 MTOPS, and that is forecast to rise to well over 10,000
MTOPS this year.
The average level for a single processor this year is 655
MTOPS; in other words, a single processor is now over the level
of what we considered a supercomputer 5 years ago, and that
level is forecast to rise to 1,135 next year, and the level of
the highest-performing single processor right now is higher
than 1,135 MTOPS.
I make these points, Mr. Chairman, to illustrate, first of
all, the rapid pace of technological development in this field
and, second, to point out that this is the real issue when it
comes to the availability and the ubiquity of this technology.
It is not primarily a question of whether the Indians or the
Russians or the Chinese are developing indigenous computers on
their own. What they are doing is demonstrating what everybody
else in the world was demonstrating, which is their ability to
string lots of small computers together or to buy commodity
chips and build advanced computers through upgrades and through
parallel processing. This is, from the standpoint of people who
are concerned about this technology getting out there, not good
news, but the fact is it is out there and, in fact, it has been
out there for a good while.
Our regulations and our policy prohibit the export under a
license exception, which means without advance approval by the
government, of computers that the exporter knows will be used
to enhance computational power above the eligibility limits
allowed for particular countries.
Beyond that, controlling computers today with complete
effectiveness would really mean individually licensing
computers down to the level of those in your office, which
would be absurd administratively and would be disastrous
economically for this sector and all of the secondary and
tertiary sectors of our economy that depend upon it.
The President's policy is a reflection of the reality of
computer technology today. It is available abroad and is
rapidly increasing in power and speed. Controls on all but the
highest levels have limited utility, and efforts to control at
lower levels will not only be unsuccessful, they will limit our
ability to widely disseminate American standards, American
software, and American hardware and, thereby, damage our
companies economically.
There has already been considerable consolidation within
the high-performance computing industry. These companies depend
significantly on their exports for their survival, and their
survival, I would argue, is essential not only to our economic
health and growth, but also to our own national security.
In the area of regulatory reform, for the first time in
over 40 years, we clarified and simplified our regulations
through a comprehensive rewrite that made them more user
friendly and easier to enforce. All of that was in line with
the goals set by the Trade Promotion Coordinating Committee,
which is an organization that the Congress authorized,
actually, while I was still here in 1992.
With respect to multilateral cooperation, the
administration has worked hard to establish the Wassenaar
Arrangement, which deals with multilateral controls on exports
of conventional arms and sensitive dual use equipment. This is
a particularly important development as we transition from
East-West Cold War controls to a regime that focuses upon
transfers of equipment and technology that could enhance
conventional military capabilities in destabilizing ways or
increase the access of rogue nations to weapons of mass
destruction or the means to deliver them.
I also want to note, Mr. Chairman, that one of the most
important things I think we have done and one of the least
heralded is our work with many of the newly independent States
of the former USSR and Central and Eastern Europe to help them
develop effective export control systems.
These initiatives are particularly important since many of
these countries possess strong technical capabilities which can
support weapons proliferation programs. They've got a lot of
equipment, they've got a lot of scientists, but they don't have
a lot of experience with border controls. And what we have
done, what Defense has done, what the Customs Service has done,
what the State Department has done is to work with these
countries to train their personnel to supply them with the
equipment that they need to maintain competent export control
systems, and in some cases to help them write their laws and
also to help them write their regulations.
This is a slow process, but the result is countries that
are newly independent are developing competent export control
systems for themselves and are learning to appreciate the
importance of those systems, particularly if they want to
ultimately join the Wassenaar Arrangement, as some of them
have--Russia and Ukraine--or other multilateral regimes that
are out there.
We think this is a very important program which,
incidentally, was funded with Nunn-Lugar money through the
appropriation and authorization proves that the Congress
undertook in past years. We think it is clearly in our national
interest to work closely with these countries and to help them
develop these procedures.
In all of these initiatives, I also want to mention, Mr.
Chairman, in view of your reference to a couple of cases, our
enforcement program at the Bureau of Export Administration
plays a key role, particularly as we focus more on specific
end-users and end-uses, and I am glad to point out that
Congress has supported these efforts through additional funds,
particularly in view of your status as a senior member of the
Appropriations Committee. I always like to have the opportunity
to talk to you about how important enforcement is and how
important adequate resources to do it are.
We have, in recent years, undertaken the criminal
prosecutions of persons who illegally exported zirconium for
Iraqi munitions, unlicensed equipment for India's missile
program, brokerage services for Iraqi rocket fuel, and gas
masks to suspected Aum Shinrikyo terrorists in Japan, just to
name a few.
These investigations also included the first civil charges
and penalties for alleged unlicensed exports of biotoxins which
are controlled to prevent proliferation. Just 2 weeks ago we
executed a search warrant on a firm that apparently shipped
software for integrated circuit design to China without the
proper license.
BXA prohibits export of items that would make material
contributions to proliferation projects abroad, regardless of
whether such items are specifically listed on our Control List
in our regulations. Using the Enhanced Proliferation Control
Initiative provisions of our regs, an exporter must apply for a
license when he or she knows or is informed by BXA that the
end-use of an item may be destined for a project or activity of
proliferation concern. Our regs also prohibit any person from
supporting proliferation projects in any way, even when there
are no U.S. products or no export transactions involved.
As an example, following a joint investigation of several
agencies, a Long Island resident pled guilty to violating the
EPCI provisions of our regulations by brokering the sale of
Chinese-origin ammonium perchlorate, which is a rocket fuel
ingredient, for shipment to Iraq. The shipment was stopped. The
individual was apprehended.
In order to save a little time, let me skip some comments
in my formal remarks, Mr. Chairman, and make a few brief
comments on Russia and China, which I know are countries of
concern to you.
Russia is continuing to develop its own export control
system and is in the early stages of participating in
International Export Control Regimes. As I mentioned, it is a
member of Wassenaar. It is a party to major nonproliferation
treaties and agreements. It has signed but not yet ratified the
Chemical Weapons Convention.
We are encouraged by these developments and hopeful that
they will enable us to work out problems in a cooperative way,
including cases of diversion or illegal purchases.
At the same time, as Mr. Einhorn reported to you last week,
although Russian policies with respect to the development and
export of weapons of mass destruction are encouraging, actual
events from time to time are not consistent with those
policies. Until we see greater consistency between Russian
policy and practice, including a Russian export control system
that is more reliable and fully harmonized with our own--and we
are working with them on that and that of our other Wassenaar
partners--we will continue to maintain appropriate controls on
exports to Russia.
Finally, on China, let me close with a brief note. The
administration policy toward China, as you know, is one of
constructive engagement. We seek to engage with China to
strengthen cooperation in areas where we agree and resolve
differences where we do not.
Our overall goal is to encourage China to become integrated
into the world system and to meet international norms of
behavior in nonproliferation, in export controls, as well as
other areas. We believe that expanding trade, business,
academic, and government contacts with China supports this
goal.
The administration rejects the view that China is an enemy
that must be contained. Our export control policy toward China
seeks to support our engagement strategy and the creation of
higher paying export-based jobs in the United States while
denying licenses for items whose export would pose significant
national security risks to the United States.
For this reason, the vast majority of U.S. exports to China
proceed with no objections by the U.S. Government. However, we
scrutinize carefully exports which might raise national
security concerns. We also continue to maintain Tiananmen
Square sanctions, which limit the items that can be licensed
for China. Where appropriate, we impose sanctions on Chinese
entities for proliferation or other activities consistent with
U.S. law.
I would also note in passing, Mr. Chairman, as an example
of the scrutiny that we provide, the licensing data that we
have for China suggests that over the last several years the
denial rate has tripled compared to previous years. We are
spending a great deal of time on Chinese license applications
and examining them very carefully.
We are proud in this administration of our Strategic Trade
and Non-proliferation record. We think we have developed an
effective interagency process that facilitates legitimate
trade, while restricting transfers that are inimical to our
national interests. We have strengthened our enforcement
capabilities with your support, and we have worked effectively
with the business community to enlist their support for our
control initiatives, which is absolutely critical.
In the years ahead, we will continue to try to do exactly
those same things. We also look forward to working with the
Congress more closely even than we have to those same ends.
Thank you.
[The prepared statement of Mr. Reinsch follows:]
PREPARED STATEMENT OF UNDER SECRETARY WILLIAM A. REINSCH
Senator Cochran. Thank you very much for your comments and
for your statement, which we have put in the record in full.
We will now go to Dr. Wallerstein for his comments. We have
a copy of your statement, Dr. Wallerstein, and we will put that
in the record in full and encourage you to make such summary
comments as you think would be helpful to the Subcommittee.
You may proceed.
TESTIMONY OF MITCHEL B. WALLERSTEIN, DEPUTY ASSISTANT SECRETARY
FOR COUNTERPROLIFERATION, DEPARTMENT OF DEFENSE
Mr. Wallerstein. Thank you, Mr. Chairman. I am pleased to
do so.
Secretary of Defense Cohen stated in his 1997 annual report
that technology security and export controls are an important
element in strengthening the preventive defense pillar of U.S.
defense strategy. Secretary Cohen emphasized that DOD's
technology security efforts serve two main purposes; first,
they seek to prevent the proliferation of nuclear, biological,
and chemical weapons--what we refer to as NBC weapons--and
their means of delivery, which are primarily ballistic and
cruise missiles.
Secondly, export controls seek to preserve U.S. military
technological advantages by controlling conventional arms and
sensitive dual-use goods, services, and technology.
Proliferation threatens U.S. national security interests.
There is certainly no question about that. It can exacerbate
regional instabilities and increase the threats to U.S.
interests worldwide, particularly in regions where we may be
likely to deploy forces, such as Northeast Asia and the Persian
Gulf.
DoD believes that this proliferation threat can be
effectively addressed through support for nonproliferation
regimes, promotion of effective national export controls, and
close export control cooperation with foreign governments that
are responsible members of the world community and that share
our concerns regarding proliferation.
We know that carefully targeted and rigorously enforced
export controls can and do dramatically slow the pace of
proliferation and raise the cost to potential proliferators.
We also believe that it is important to continue to
carefully regulate exports of potentially destabilizing
conventional arms and sensitive dual-use technologies. It is no
coincidence that the countries seeking NBC weapons and missile
delivery systems are also simultaneously attempting to build up
their conventional weapons capabilities.
Let me also note that the Department of Defense sees no
signs that the underlying forces, which are causing NBC weapons
proliferation and destabilizing conventional arms build-ups is
abating. The post-Cold War era is characterized by global
diffusion of technology and increasing indigenous expertise,
which contributes to more widespread production of high
technology goods in many regions.
That production, in turn, makes possible the application of
advanced civilian technologies to military users.
DoD has special responsibility to provide our Armed Forces
with the best and most technologically advanced equipment for
fighting future conflicts and for protecting the safety of
these men and women. Our fighting men and women performed
brilliantly in Desert Storm in large measure because they had
the most advanced technology, which they needed to maintain
conventional superiority on the battlefield.
We must continue to provide the most advanced equipment to
our fighting forces and ensure that this equipment is superior
to that of any foe. Export controls are essential in
maintaining our technology lead in key military systems.
Let me emphasize a few major principles that I believe
should be kept in mind in implementing export controls.
First, is the need for a strong policy on which to control
and, as required, to impose conditions or to deny sensitive
exports to any destination for reasons of national security or
foreign policy.
Second is the need to retain substantial administration
flexibility in both establishing and implementing controls.
And third is the need to maintain a sufficiently broad
basis for imposing unilateral controls under certain limited
conditions, while we endeavor at the same time to make such
controls more fully multilateral in their impact.
I believe that we have already moved effectively to
implement these principles by improving the efficiency and the
transparency of the U.S. Government export control process.
As Under Secretary Reinsch has noted, in a recent Executive
Order, the President has directed that there will be
appropriate interagency review of all dual-use export licenses,
thereby addressing a Congressional concern that the Department
of Defense has on occasion in the past not been afforded the
opportunity to review certain dual-use exports.
The Executive Order also imposes, as Under Secretary
Reinsch said, rigorous time constraints that allow us to
account for national security concerns while still providing
for expeditious review of license applications.
The new Executive Order is an example of this
administration's efforts to streamline the export control
process, to tighten controls where necessary, but still to
ensure that U.S. exporters remain competitive in the world
market.
At the same time, multilateral export control frameworks
have been enhanced by the establishment of theWassenaar
Arrangement on export controls for conventional arms and dual-
use goods and technologies. This Wassenaar Arrangement--which,
by the way, the name derives from the town outside The Hague in
the Netherlands where the agreement was negotiated--compliments
other existing multilateral nonproliferation regimes
specifically directed at curtailing the spread of weapons of
mass destruction and their means of delivery. This includes the
Nuclear Suppliers Group, the Missile Technology Control Regime
and the Australia Group.
In sum, Mr. Chairman, proliferation is a multifaceted
challenge that spans the full spectrum of conflict and
threatens peace and stability at different levels--globally as
well as regionally. It is not a challenge that will soon go
away. For this reason, it is appropriate and necessary to use a
wide range of national and international resources, including
effective export controls, in our attempts to control the
proliferation of nuclear, biological, and chemical weapons,
missile delivery systems, as well as the dual-use goods and
technologies that contribute to them.
I will conclude my statement there and am happy to respond
to your questions.
[The prepared statement of Mr. Wallerstein follows:]
PREPARED STATEMENT DR. MITCHEL B. WALLERSTEIN
Senator Cochran. Thank you very much, Mr. Secretary.
Before proceeding with questions, I want to welcome our
good friend from Illinois, Senator Durbin, and yield to him for
any opening comments that you would like to make.
Senator Durbin. I will just ask questions later.
Senator Cochran. Secretary Reinsch, in your comments that
you made to the Committee, you talked about the fact that in
this day of emerging technologies that they are much
moreadvanced today than they were even 4 years ago and,
particularly, in computers, that there already is out there the
capability to develop computers with the power that we used to
call supercomputers that are now ordinary, everyday computers.
But isn't it a fact that only the U.S. and Japan are the
manufacturers who are capable of manufacturing the true
supercomputers in today's jargon?
Mr. Reinsch. I would like to say that we cornered the
market on that, Mr. Chairman, because I think that would be
good news, and we have a study underway to determine the answer
to that question.
Right now, based on the information available, I would say
that is, by and large, correct, but it misses the point. As I
said in my statement the real issue is upgrades, parallel
processing, the ability to assemble computer power through work
stations and the acquisition of single and multiple processors
that are uncontrolled, and widely available, and those are
widely produced in lots of other countries.
Senator Cochran. A General Accounting Office review of
computer export data indicates that it is unlikely that Russian
military and nuclear weapons laboratories had acquired
computers capable of more than approximately 3,500 MTOPS--
million of theoretical operations per second--due to a lack of
known sales of computers above that capability from the United
States or Japan, and then they say these are the only countries
currently producing computers above that level.
Is that a correct statement? Is GAO right about that; that
the United States and Japan are the only countries currently
producing computers above the 3,500 MTOPS level?
Mr. Reinsch. I cannot, at this point, make a convincing
case that that is wrong, Mr. Chairman. There was testimony on
the House side on this point by Ken Flamm, formerly of the
Defense Department and now at the Brookings Institution, that
suggested that there were some other producers, but I don't
have that information, and I am not prepared to put it forward.
For purposes of this discussion, I am happy simply to assume
that that is correct.
The question, of course, really is, though, what would the
Russians have done or what would they have been able to do had
an American company not sold them the computers that are at
issue? Would they have been able to obtain comparable computing
power through other means. I think the answer to that is yes.
They didn't have to go down that road because the sale took
place.
Senator Cochran. The policy that this administration now
has, as I understand it, classifies different countries to
which U.S. manufacturers are permitted to export comuters of
certain capabilities. There are Computer Tier I countries,
including Western Europe, Japan, Canada, Mexico, Australia, New
Zealand. No license is required for supercomputer exports to or
re-exports among those countries. So that is a license-free
zone that we have described, as I understand it.
There is a second tier in the new policy, which includes
South America, South Korea, the ASEAN nations, Hungary, Poland,
the Czech Republic and others, where no license is required to
export supercomputers with capabilities up to 10,000 MTOPS.
Record keeping and reporting by the manufacturer, though, is
required.
And then there are the Tier III countries, and those are
the ones of particular concern that we are talking about today,
comprised of India, Pakistan, all of the Middle East not
included in other tiers, states of the Former Soviet Union,
China, Vietnam, and the rest of Eastern Europe. Export
requirements under this Tier III licensing requirements are
somewhat complicated depending on who the end user is, military
or civilian, and what the end use is, military or civilian. And
the license applications, required, as I understand it, for
these countries are supposed to be examined on a case-by-case
basis, and these individual validated export licenses are
required to export to or re-export among Tier III countries
computers capable of greater than 2,000 MTOPS to military end
users or end uses in these countries. This, of course, includes
nuclear, biological, chemical, or missile-related end-uses.
Is that a fair characterization of our policy and the
regulations that your office is responsible for enforcing?
Mr. Reinsch. Yes, it is, Mr. Chairman.
The only minor point I would make is with respect to Tier
III and the Middle East. It is the non-embargoed Middle East.
There is a Tier IV, which includes a number of countries in the
Middle East, like Iran and Iraq, where the effective limit is
six MTOPS, and it hasn't changed in years.
Senator Cochran. Right. Cuba, Iraq, Iran, Libya, North
Korea, Sudan, and Syria.
Mr. Reinsch. Yes. They are in a sep----
Senator Cochran. There is another tier, Tier IV.
Mr. Reinsch. Yes. I wouldn't want anyone to think that they
are in Tier III.
Senator Cochran. That is the embargo. No supercomputer
sales are permitted to those Tier IV countries; is that
correct?
Mr. Reinsch. Well, effectively. Our limit is six MTOPS,
which eliminates everything.
Senator Cochran. That is a typewriter, isn't it?
[Laughter.]
Mr. Reinsch. Yes. Approximately.
Your description of the Tier III policy, as I heard it, is
correct.
Senator Cochran. Let me ask you this. If I am an exporter
and I want to sell a supercomputer to one of these Tier III
countries, China, for example, is there any way I can consult
and get a list of suspected end users that would be prohibited
under this policy? Can I consult with you so you can give me a
list of those that I shouldn't sell to in China, for example,
who are military end users?
Mr. Reinsch. There are approximately three things you can
do. You can always consult with us, and people do that.
Normally, that consultation takes the form of a company coming
in and saying, ``We intend to do business with X. Is that OK?
What do you think of X? Do you have information about Entity X,
whatever it is? Is that an end user that would require a
license?''
We are prepared and have told the companies that we are
prepared to answer those questions when they come in. We had a
meeting with--this is a fairly small universe of producers, by
the way, six or seven--and we had a meeting with them shortly
after this policy became effective and went over the procedures
that we wanted them to follow, the kinds of records that we
wanted them to keep, which they have been keeping, and the
opportunities they had to consult. They can come in and do
that.
In addition, second, we can inform them individually or
collectively of end users that are problematical within the
meaning of the President's policy or, alternatively, third, we
can publish in the Federal Register the names of entities that
we have identified as proliferation end users for which a
license would be required.
We have begun to do that. We have not done it extensively
so far. There are intelligence sources and methods issues that
come up frequently, as well as some other considerations.
We have thus far published--and this is a policy that we
adopted last year in terms of a mechanism for working these
things through and making a decision--thus far we have
published two names, and we expect, within the next week or so
to publish a significantly longer list of more names that will
include Chinese names. The two names that we published were in
Israel and India.
Senator Cochran. It is my understanding that the Department
of Commerce has refused up to now to make available any listing
of military users in Russia or in China or in India or in
Pakistan and that the only one, when asked, that was identified
was in Israel. Is this sort of trying to shut the door after
everything is already out?
Mr. Reinsch. This is a frustrating question, which I know
has been the subject of some comment in the newspapers. The
reason I am frustrated, frankly, Mr. Chairman, is because you
are sitting here talking to the two people that have been
trying to get this information out and publish this information
for a long time.
The decision to publish information, however, is not one
that resides exclusively in a single agency. This is an
interagency decision. As I said, with virtually all of these
matters there are sources and methods, and intelligence-related
questions that have to be debated and considered, and sometimes
we don't publish, frequently we don't publish for that reason,
even though we have identified someone that, for other reasons,
ought to be published.
Senator Cochran. The practical result has been to put the
exporters on the honor system and to give them the
responsibility for determining who is a military end user or
what will be a likely military end use.
Mr. Reinsch. I don't agree with that, Mr. Chairman. I think
that overstates it.
As I said, we have had a good bit of consultation with
them, talking to them about what to look for, red flags, what
kind of indicators they ought to identify in their customer
business. We have invited them to come in and consult with us
regularly. In the Russian situation, and I assume you are
familiar with the facts of that, most of the companies in this
universe of companies did come in and consult with us about the
Russian end users and, in fact, they submitted licenses
acknowledging that these were end users for which licenses
needed to be submitted. We didn't publish those names, but they
figured it out. It wasn't very hard to figure it out. They
consulted with us.
We declined to approve those licenses. They got the
message. There is a company that you mentioned that did not get
that message, apparently, and undertook the sale, and that is a
matter of investigation right now with the Justice Department,
and I can't comment further on the case.
But I would say that, by and large, these companies have
not had a lot of difficulty figuring out who the military end
users are and which are not.
Senator Cochran. There was some statement in your remarks
about how many supercomputers have been purchased under this
new policy by China. I think 46 is the number that I remember.
You may have mentioned that in your testimony over on the House
side at a hearing there.
Mr. Reinsch. That is what I said.
Senator Cochran. You told us a specific number of
supercomputers that have been purchased in Russia and China.
How do you know there aren't more than that in those two
countries? How do you know there are just 46 in China, for
example, and do you know where they all are?
Mr. Reinsch. The companies, under the President's policy,
are required to keep records of all sales worldwide. They have
done so. They have submitted those records to us. The numbers
that I cited in that testimony were the numbers that we had
available at that time that the companies had represented to us
were the sum total of their sales.
Now, as with anything in life, there are two possibilities;
one, they may have forgotten something, and in point of fact,
we have got some additional ones dribbling in, only one
additional one for China. But there is always that possibility
as they go through their records and recalculate.
There is also the possibility, of course, that they are
lying to us; that someone is engaging in fraud. That is an
enforcement matter. That is why I have enforcement agents who
do a variety of things that I would prefer not to get into
publicly to test the validity of the information that we're
given and to work with parties other than the companies
themselves on those points.
We don't simply take their word for it. But that is where
we begin.
Senator Cochran. Let me ask you this about the end uses to
which the supercomputers have been put. Are you satisfied that
none of the supercomputers have been used to upgrade the
quality of nuclear weapons in China?
Mr. Reinsch. We have no evidence that any of them have been
used for that purpose. We have a very high level of confidence
on that point with respect to all but two, based on the kind of
end user it is, and there are a couple where we are looking
into the matter further, but that is not based on any evidence
that there is a problem. It is based on our desire to learn a
little bit more about the nature of the end user.
Senator Cochran. There have been published reports in the
press that the Chinese Academy of Sciences is involved in
assisting in the upgrade of nuclear weapons capability or
missile technology in China. Do you agree with those press
reports?
Mr. Reinsch. Mr. Chairman, that is something that the
intelligence community has looked into in considerable detail.
We have information on that, but it is classified and I can't
provide it to you in open session.
Senator Cochran. Do you have any evidence that any of the
supercomputers, which have been sold by U.S. firms have
violated your export control policies that have not been
reported in the press?
Mr. Reinsch. Well, I can't keep track of everything the
press reports. There are three cases that are under
investigation; the two in Russia, which have been reported in
the press, and the single one that you alluded to with respect
to China and the Chinese Academy of Sciences, which has also
been reported in the press, is one that is being looked at.
Senator Cochran. And that is the one that is involving the
Chinese Academy of Sciences?
Mr. Reinsch. That is correct. That is the one that was
stated in the press.
Senator Cochran. Do you know whether Silicon Graphics has
sold any high-performance computers to countries that are
proliferation risks other than Russia and China?
Mr. Reinsch. We have their complete records, Mr. Chairman.
I would have to look it up. I was focused on China for this
hearing. We can find out.
Senator Cochran. I would appreciate your providing that for
the record, if you could.
Does the Commerce Department have a list of the 1,100 high-
performance computers which documents the end user, the speed
of the computer, the date of export, value and the identity of
the exporter?
Mr. Reinsch. Yes.
Senator Cochran. Could you furnish that to the Committee
for our record?
Mr. Reinsch. I was afraid you were going to ask me that,
Mr. Chairman. This is information that is protected by Section
12(c) of the Export Administration Act. Section 12 of the
Export Administration Act requires us to provide this
information to the Congressional Committees of appropriate
jurisdiction and prohibits them from making that information
further available except by a vote of the full Committee.
We have not yet made a judgment as to whether this is a
Committee of appropriate jurisdiction. The main Committee of
appropriate jurisdiction in the Senate is the Senate Banking
Committee, which has not requested this information. I would
have to consult with my lawyers, frankly. We have not had a
request from your Committee before and haven't made a judgment
on whether you fall within the meaning of 12(c).
Senator Cochran. Since the Nuclear Nonproliferation Treaty
was signed, this Committee has had the responsibility of
oversight of compliance with the terms of that agreement, and
we annually review the status of that and the compliance with
the treaty terms by signatories. I can recall when I first
became Chairman of the Subcommittee that had jurisdiction over
that subject, I met regularly with the ambassador, who is our
delegate to the Vienna IAEA Conference on the subject of
safeguards and compliance with safeguards.
Senator Chuck Percy had the responsibility of chairing this
Subcommittee at one time. Other Senators have as well. Senator
Scoop Jackson at one time had responsibilities with respect to
this subject. The Committee continues to exercise jurisdiction
over this area of proliferation, and that is the responsibility
that we are undertaking to discharge in the conduct of these
hearings. So I think it is clearly established that here in the
Senate the Subcommittee is the Committee of jurisdiction.
Having said that, I would be glad to take it up with the
Chairman of the Committee and other members of the Committee
for further discussion. But if it is determined that I am right
about that, we will resubmit that question in writing and ask
for you to produce that information. But we will be glad to
explore that further. I respect your position that you are
taking at this point.
Mr. Reinsch. You make a very compelling case, Mr. Chairman.
I hope you can appreciate the situation that I am in. We have
no reluctance to provide the information to the Congress, and
we have told the Committees that clearly our--for example, our
authorizing Committees who made a similar request--that we are
happy to provide the information. So we don't have any problem
with it coming to the Congress.
I have to defend the law as it's written, and I have to
consult with my lawyers, but I certainly understand the
strength of your case, and I understand, also, the very broad
jurisdiction the Government Affairs Committee has.
Senator Cochran. Dr. Wallerstein, you mentioned that the
Defense Department has concerns in this area and
responsibilities as well, and with the interagency guidelines
that have now been promulgated I assume that part of your
responsibility is to assess the security risk of the exports of
these over a thousand high performance computers that we know
have already taken place.
Have you come to any conclusion about whether or not these
exports do pose a new security risk to the United States?
Mr. Wallerstein. Senator, we review each of these proposed
exports on a case-by-case basis and provide our views back to
the Commerce Department, in the case of dual-use licensing and
to the State Department, in the case of licensing of munitions
exports.
In some cases we recommend conditions be imposed on the
exports; and that can be done, particularly for machines of
higher capability. So that we may have a higher level of
confidence that the machine is being used for the purposes that
are proposed in the export license.
Based on our case-by-case assessment and on the
conditionality that we have on occasion recommended, and that
has been implemented, we have no immediate evidence to suggest
that the exports to China or to any other country have been
inimical to U.S. national security interests.
Senator Cochran. It is my understanding that the
capabilities of these supercomputers are such that they can be
used and may have been used to develop smaller nuclear warheads
for missiles and to improve the accuracy of missiles that are
used to deliver weapons of mass destruction.
Do you have any evidence to support the conclusion that
some of these computers have been used in those ways?
Mr. Wallerstein. Sir, I have no immediate evidence to
document the assertion that you are making. That said, I would
certainly acknowledge that, as Under Secretary Reinsch has
already indicated, with the global diffusion of computing
technology, there is wider access to more capable computers.
I am sure you have heard the assertion made that the
original designs for the first U.S. nuclear weapons were done
on slide rules or on very primitive calculating machines. There
is no question that lower-powered computers can aid in certain
kinds of military applications, but those computers have become
commodities at this point in time in 1997.
What we have determined, and what was integral to the
change in policy that was implemented in 1995, were that there
were applications that were well above the levels that we
permit to be exported without a validated license that are
essential to U.S. national security, and we have safeguarded
those applications.
Senator Cochran. Do you know whether our government, either
the military or other agencies of our government, has conducted
any analytical studies to try to assess the threat these
supercomputers could present to our military or to our national
security before concurring with the administration's assessment
of essentially decontrolling as a matter of national interest
the sale of supercomputers above the 2000 MTOPS threshold?
Mr. Wallerstein. Senator, I have direct responsibility in
the Department of Defense for this matter, and I can assure you
that both in the 1993 computer policy change and, again, in the
1995 computer policy change, all elements of the Department of
Defense were integrally involved. This included all parts of
the Office of the Secretary of Defense as well as the Joint
Chiefs of Staff and the services.
Senator Cochran. I have some more questions about China,
and Russia, and also Iran, but I am going to defer to my
friend, Senator Durbin, for any questions he might have at this
point.
Senator Durbin.
Senator Durbin. Thank you, Mr. Chairman. And thank you for
this hearing. I think it is an important and fascinating topic.
I want to try to come to grips with an understanding about
the current export policy that this administration has
instituted. One of the critics of that policy has said that we
have--I won't use his words, but I will say we have reduced
export controls on strategic technology to one-tenth of what
they were under the Bush administration. Under President Bush,
no computer performing more than 12.5 million operations per
second could be sold to Russia or China without a license. Now
computers up to seven billion operations per second can go
without a license if the sale is not to a nuclear, chemical
missile or military site.
Is that a fair characterization of the change in export
policy?
Mr. Reinsch. Yes. Strictly in terms of theoretical level of
performance, yes.
Senator Durbin. And, of course, then it raises the question
if we are turning loose this level of technology, this
expanding level of technology, what controls do we retain in
that process? I think that is what this hearing is all about.
I am troubled by some of the things that have been said.
The whole question of dual use assumes, does it not, some
cooperation on the part of the purchaser in terms of end use
and disclosure of that end use?
Mr. Reinsch. Well, to grant a license the end user has to
be identified, and we can make judgments about the nature of
that end user. Those are, in part, made based on
representations that the end user as well as the American or
exporter applicant might make. But the licensing process is
also informed by intelligence information, enforcement
information, other things that we know about the end user that
he or she may not tell us or may not want us to know.
Senator Durbin. So we try, when we don't trust the
purchaser, to verify the end use through our own surveillance
within that country?
Mr. Reinsch. We have a variety of means. Prelicense checks
is one of them, in which our posts abroad engage in, well,
exactly what I said, a prelicense check of the facility. There
have been a couple famous cases in the distant past in the
1980s, where people have gone out after the fact to look for
the computer and discovered an empty building and the computer
had been shipped off somewhere else or discovered that the
company was a mail drop. There is a lot you can discover with
prelicense checks. There is a lot you can discover just by
wandering around a plant to determine the nature of their real
business, which they may or may not want to tell you.
The fact is, at the same time, though, it is a reality that
all exports, when they leave our shore, go into somebody else's
hands, and we don't have control over them any more. It may be
somebody in the UK, and you have a high level of confidence in
what is going to happen. It may be somebody in China and you
have a lower level of confidence about what is going to happen,
but it is equally out of our control.
Senator Durbin. Someone said earlier this century--I can't
recall the exact source--that a capitalist will sell you the
rope you will use to hang him.
I am just wondering, in this instance, whether or not we
are keeping track of this rope appropriately.
Let's take one other aspect of this. Let's assume you have
a conscientious business in the United states that doesn't want
to get caught selling to someone who is going to misuse this
product. From what you have said, if they come to their
government and say, ``Give us some guidance. We would like to
know which customers to avoid in Russia or China or some of
these other nations,'' I think your testimony was that there is
a limited amount that you can tell them.
Mr. Reinsch. We can tell them a good bit. We can't tell
them probably as much as they would like, and we can't always
tell them as much as we know because sometimes we are
constrained by the way in which we obtain the information from
revealing even that we have it.
Senator Durbin. Have we thrown in the towel when it comes
to export controls? Are we assuming that there are so many
sources of this technology around the world that we might as
well let American exporters earn some money and hope that maybe
at least we'll make a few bucks off of this deal?
Mr. Reinsch. No, we haven't thrown in the towel at all,
Senator. I think the conversation is a bit skewed because we
have been talking exclusively about computers, which is a
technology that is uniquely difficult to control for the
reasons I have said.
In a number of other areas, and I think Dr. Wallerstein can
mention them, but areas like stealth technology, advanced
materials, composite materials, very sophisticated electronics,
chemical precursors, biotoxins, a whole host of things that we
control, I think our system is very effective.
Senator Durbin. But these ubiquitous computers that tend to
be----
Mr. Reinsch. That is different. It is not unique, but it is
different.
Senator Durbin. We don't seem to have much of a handle on
them. I just wondered why, under the 1995 policy as I
understand it, most supercomputers sold for civilian purposes
do not need to be licensed for export by the Federal Government
and exporters, consequently, cannot be required to track how
they are used. Is that a fair statement?
Mr. Reinsch. No. The exporters are required to keep records
of every sale of high-performance computers.
Senator Durbin. But I am talking about end use. They can
certainly give the name of the nominal purchaser, but there is
no way to track, and what you are suggesting is only through
spy techniques can we attempt verification.
Mr. Reinsch. Well, they give us both the--they know both
the name of the end user and the end use. They know why the
machine is being is being bought. That is what they tell us.
Now, if you are asking me how do they know 12 months down the
road that the machine, A, is still there and, B, is still being
used for that purpose, that gets back to the control question.
Although, actually, in the case of computers there is a way to
tell because these things, particularly the high-performance
ones, need regular service. I wouldn't want to suggest they
break frequently, but a standard part of this kind of transfer
is an ongoing service, and supply, and parts and sometimes
upgrade relationship with the vendor. So the manufacturers know
and have an ongoing relationship with the buyer most of the
time, and know whether the machine is still there and have a
pretty good idea of how it is being used.
Senator Durbin. So do we keep or does the company keep and
file with the government a log, not only of sales and
purchases, but continued maintenance and repair so that we can
see if the end use is as it was stated at the originalpurchase?
Mr. Reinsch. They keep those records. They provide them to
us on request. In this particular case, we have requested them
and are receiving them.
Senator Durbin. We have had a couple instances, have we
not, in the last few months involving Silicon Graphics that
suggests that computer sales were made in Russia and China that
were at least suspect?
Mr. Reinsch. Yes. That is, as I said, under active
investigation via the Justice Department. I don't want to go
into a lot of detail that would prejudice that outcome, but I
think that is a fair statement.
Senator Durbin. I am not going to ask you to go into it,
but I think it really tells the story about this new policy and
the fact that we have surrendered control in a lot of areas
that, for whatever reason. I don't know if it is our belief
that the world market is so rife with these computers that we
might as well get a piece of the action or whatever reason, but
we seem to have taken a new approach to this, which is very
porous and not very accountable, as I see it.
Mr. Reinsch. If I may, Senator. I am not here to tell you
that is good news. I guess I am here to tell you that that is
technological and commercial reality. The reason this is a
ubiquitous technology is because of the chips. Semiconductor
chips, single and multiprocessors, most of which now, as a
single processor, function at a higher level than the entire
computer that the Bush administration controlled, are out
there. Lots of countries make them. Lots of countries make them
in ways that will fit into American products. Upgrades are
easy. You slide another board in, more chips, and you have got
more capacity.
You can string these things together in parallel
processing. We can export 40 Pentiums and you are at 6,000/
7,000 MTOPS right there.
I can make it even worse for you. If I were the Chinese, to
be frank, I wouldn't deal with exports. I would set up a front
company in this country, buy one, and it wouldn't even be an
export, and have them do all of the computation I wanted inside
the United States and ship the data back.
Senator Durbin. I think it is curious the date of this
hearing, it's just 1963, June 10, 1963, that President Kennedy
gave a speech at American University about his vision of the
end of nuclear weapons in the world and hoped that we would
reach it and all that has transpired since, including the end
of the Cold War and a reduction in nuclear warheads. We seem to
be on the right track there.
But as we are making tangible, measurable progress at that
level, it is probably because we are stuck in the mind-set of
the 1960s and the belief that this is the protection of our
future.
It appears that the challenge for the new century is in
technology, where the right computer can provide, from what I
have read, as much or more information than nuclear testing
used to provide in years gone by.
From what I hear and your testimony, this is not
controllable. It is not a question of counting warheads. And
there is such an easy commerce in this technology that holding
out the prospect of controlling proliferation may be naive.
I don't know if our export policy makes sense. I have real
serious questions when it comes to China and certainly as to
Russia.
Would a flat prohibition on the sale of dual-use technology
to countries who refuse end-use verification be effective?
Mr. Reinsch. Not in the computer area, no, for the reasons
that you and I have both said.
Senator Durbin. They will buy it from someone else.
Mr. Reinsch. Sure. But let me not leave you--I mean, you
have made some very thoughtful observations, Senator, but let
me not leave you with as little hope as you have suggested by
observing that the computer is neither the beginning nor the
end nor the larger part of our proliferation policy. You can,
as Dr. Wallerstein said, design a nuclear weapon without one,
but even if you designed one, you need a lot of--to build a
bomb, to build a missile you need a lot of things besides a
computer.
You need a lot of special materials, beginning with uranium
or plutonium. You need to be able to have a continuous supply
of that. There are a lot of other special materials, including
special steel, that goes into the making of the bomb. A missile
has all kinds of electronic systems, special materials and
other things that are an integral part of making it function.
These things we control, and we control them very
effectively, and they are not ubiquitous technologies in the
way that computers are. So I would not, while I am gloomy in
the sense about the utility or the possibility really of
controlling a technology that is widespread with respect to
software and the intellectual [inaudible] computer you can
export over the phone. Think of the enforcement problems
associated with that. While I am gloomy about that, I am not
gloomy about our ability to deter proliferation because there
are so many other pieces of the puzzle where I think what we
are doing is very effective.
Mr. Wallerstein. Senator, let me pick up specifically on
the nuclear aspect of this, which you were just addressing.
As you know, the other part of President Kennedy's famous
statement in 1963 was that he predicted there would be over 20
nuclear capable States in the 1970's. Of course, that never
came to be; in part, because, as Under Secretary Reinsch has
indicated, we control effectively a range of technologies. We
also have participated and have increased the robustness of the
Nuclear Suppliers Group and the effectiveness of IAEA.
The other point to make here is that, while it is certainly
or it may be the case that some elements of our computer policy
have had to reflect the growing worldwide availability of
computers, the level of computational capability that is
required to run the very sophisticated models that are involved
in nuclear safety and surety are well above those that we were
talking about earlier; that is, the 7,000 MTOP threshold.
Moreover, the states that have signed the CTBT will not be
doing any further testing. And any state that would test which
is not a member we would have other means to address that, and
we are certainly not selling advanced computers to those
countries.
So I think that with respect to your concerns about nuclear
safety, and nuclear security, and nuclear nonproliferation that
our policy is in tune with those concerns and that we still
have the ability to control these higher level machines. As we
have said, we do not allow exports to military or defense end
users and we would, in any case, require a validated license.
Senator Durbin. I would be remiss if I didn't at least ask
the follow-up question if the same response would apply when it
comes to biological and chemical weapons, since we have, as you
have indicated, some elements involved in nuclear weaponry that
can and are carefully monitored. Can the same be said of the
biological and chemical weapons?
Mr. Wallerstein. Well, of course, with the advent now of
the Chemical Weapons Convention, we have a very, very large
number of countries in the world, including the United States,
which are now committed to restrict the export of chemical
precursors and other elements that are required for the
production of chemical weapons to states that are
nonsignatories.
We also have the Australia Group, which controls the export
of both chemical precursors and biological agents that are
necessary for CW and BW weapons.
I would also note, however,that these are classically dual-
use technologies. We have to remain very vigilant here because,
particularly with biological weapons, there is an ease of
concealability problem; due to the fact that these kinds of
weapons can be manufactured inside of pharmaceutical facilities
that are also producing for legitimate civilian civil end-use.
So we do need continued vigilance, but we feel that, again,
with our controls and with our multilateral commitment, through
the Australia Group, the Chemical Weapons Convention, and the
Biological Weapons Convention, we are addressing that.
Senator Durbin. Thank you. Thank you, Mr. Chairman.
Senator Cochran. Thank you, Senator.
I am told that Russia's Minister of Atomic Energy made a
surprising announcement in January that his ministry had
purchased five American supercomputers; four from Silicon
Graphics and one from IBM for two Russian nuclear weapons
design labs; Chelyabinsk-70 and ARZAMAS-16.
The minister's announcement was particularly shocking,
given the Commerce Department's decision not to approve export
license applications for similar supercomputers to the Russian
Ministry of Atomic Energy in the fall of 1996. The press
publicized this nonapproval to Hewlett-Packard and IBM.
Secretary Reinsch, if the Russian Government can obtain
from the United States without an export license supercomputers
for its nuclear weapons design labs, how can you say that the
administration's new export restrictions on high-performance
computers is serving its intended purpose?
Mr. Reinsch. Well, these are cases, as I said, that are
under investigation.
Senator Cochran. This is the first time I have mentioned
these. You hadn't responded to this question before.
Mr. Reinsch. I mentioned the Russian cases before. This is
a situation in which we thought we had done a very effective
job publicly and privately. I have to be careful because, as I
said, there is a Justice Department investigation going on
here, Mr. Chairman, and I don't want to interfere with it.
We thought we had done a very effective job publicly and
privately in indicating to the companies in this small universe
what was appropriate with respect to those institutions and
what was not. We have the obvious fact that one or more
companies didn't get that message. I think that will play out
in the criminal justice system. I don't think that is a
question of policy, frankly.
Senator Cochran. So this involves Silicon Graphics as well
then.
Mr. Reinsch. Yes, as reported.
Senator Cochran. You mentioned that the Justice Department
was investigating Silicon Graphics.
Mr. Reinsch. As publicly reported, the sales to Russia were
units by Silicon Graphics and one by IBM, and those are the
investigations that are underway right now. They have both been
referred to the appropriate Assistant U.S. Attorney.
Senator Cochran. There was published by the Department a
Russian Defense Business Directory indicating Russia's military
sites in order to acquaint potential exporters of the fact that
they shouldn't export or they should obtain permission to
export before they sold anything. Why is it that you published
the Russian Defense Business Directory to acquaint people with
potential illegal or improper purchasers but didn't publish a
similar guide for China? Was there any reason for that?
Mr. Reinsch. Yes. That wasn't the purpose of the directory,
Mr. Chairman. That directory was funded by Nunn-Lugar funds,
and our activities in this area were restricted to Russia,
Ukraine, Kazakstan, and Belarus under the Nunn-Lugar
formulation. The purpose of that directory--and we publish
directories for several of those other States, too--was to
assist in defense conversion in those countries; that is,
trying to get Russian or Belarussian or Ukrainian or whatever
defense companies out of the missile or defense or weapons
business and into other businesses, and we were trying to help
American companies understand what kind of capacity there was
over there for joint ventures or other kinds of trade or deal-
making in civilian areas. That was the purpose of the
directories.
Senator Cochran. The Russian minister, Mikhailov, made it
clear that Russia intended to use the supercomputers to design
new nuclear weapons. If the Defense Department or if Commerce
had known this at the time, would it have supported an export
license request to sell computers for that purpose, Secretary
Wallerstein?
Mr. Wallerstein. Senator, our policy is clear and
unequivocal. We do not support the export of any computers that
would assist the Russian nuclear weapons design, safety or
surety program.
Mr. Reinsch. And as you noted we did not support it when we
were presented with applications to send similar machines to
the same places 3 months earlier.
Senator Cochran. I have been told that Silicon Graphics has
indicated an intention to upgrade the computer that it sold to
the Chinese Academy of Sciences. If that is true, Dr.
Wallerstein, is it your opinion that the Department of Defense
would object to any license for that purpose?
Mr. Wallerstein. Senator, I would have to see the details
of the specific proposal and refer to our technical experts
within the Department to determine the nature of that
application and whether we could agree to that license, or if
we would require additional conditions on the license.
Senator Cochran. Secretary Reinsch, do you know whether
Silicon Graphics has actually made application for a license to
upgrade the computer that it sold to the Chinese Academy of
Sciences?
Mr. Reinsch. Not that I am aware of, Mr. Chairman. As far
as we know, no, they haven't. I would like to say, just to go
back to something I said very early on this point, Mr.
Chairman, it might be fruitful for us, and I think I can do
this because it is not 12(c) information, although it is
classified, we might want to find a way to share with you our
evaluation of the Chinese Academy of Sciences. You might be
interested in that.
Senator Cochran. We may very well do that then.
With your consent to appear, we can make available a time
that is mutually convenient for the Committee and for both of
you, and we can hear that in a classified session. We have done
that with other witnesses on other subjects as a part of this
series, and I think that is a good idea, for us to have the
full story.
Let me just say, to further elaborate on this issue about
the jurisdiction of the Committee, I omitted to say that
Senator Glenn, of course, has been Chairman of this Committee,
too, and this Subcommittee as well. And in a letter from Acting
Under Secretary Barry Carter, February 16, 1994, providing
information on the nuclear referral list and information that
had been requested in a letter from Senator Glenn, he says this
on page 2 of his letter, ``We are providing this licensing
information to you as the Chairman of the Committee on
Governmental Affairs pursuant to the confidential provisions of
Section 12(c) of the Export Administration Act of 1979, as
amended.''
We are that Committee. That is the Committee.
Mr. Reinsch. It sounds like my predecessor already made
that judgment, which is nice to know.
Senator Cochran. Yes. So there is precedent.
Mr. Reinsch. I wouldn't want to be inconsistent with my
predecessor, particularly because of the respect I have for
him, Mr. Chairman.
Senator Cochran. I am just saying this for the benefit of
your lawyers, who didn't particularly do much research, I
think.
Mr. Reinsch. No. My lawyers have not given me an opinion.
Senator Cochran. Oh. Oh, I thought you said your lawyers
have cautioned you about giving us----
Mr. Reinsch. They have not given me one, and I will make
sure they are apprised of this. The only distinction I would
make is that there is--and we have made this distinction on the
House side with respect to this same material--there is a
distinction between the full Committee and the Subcommittee,
which probably doesn't make any practical difference who signs
the letter.
Senator Cochran. I said I was going to ask you a question
about Iran, and I am.
The issue that I want to ask you to tell me your views
about involves the possibility of using this new policy to sell
supercomputers, 7,000 MTOPS, to a country in the Middle East,
who could then transfer the equipment or make the sale to Iran.
For example, the United Arab Emirates is a Tier III country
under the administration's policies on decontrolling U.S. high-
performance computer sales. U.S. supercomputer manufacturers,
then, could sell to buyers in Dubai without an export license,
providing it's a civilian buyer for civilian use up to 7,000
MTOPS. It is a fact that Iran imports more goods through Dubai
than through its own ports because of Dubai's trans-ship to
Iran. There is nothing to prevent the supercomputers from going
on to Iran or anywhere else, for that matter from Dubai.
The General Accounting Office found in 1994 that the
Commerce Department system of post-shipment verifications was
ineffective. Dr. Wallerstein, last week Deputy Assistant
Secretary of State Einhorn expressed his concern in testimony
before this Subcommittee about Iran's ongoing pursuit of
weapons of mass destruction and ballistic missile delivery
systems.
Does the Defense Department share this concern and do you
also share the concern that there is inadequate safeguard to
prevent trans-shipment of supercomputers under these new
policies?
Mr. Wallerstein. Senator, we certainly do share the State
Department's concern about the general pattern that we see
emerging with respect to Iran and its attempts to acquire
nuclear, biological, chemical weapons and missile delivery
capability. We see a widespread and fairly sophisticated effort
underway to evade the controls that are in place, and we are in
regular contact--we, that is, the U.S. Government is in regular
contact with our key allies and other major exporters to try to
assure that they will not be successful.
Certainly, as we have already indicated in this hearing,
there is a lot of material out there in world markets,
particularly in the computing area, which is beyond control. So
there is no way that this can be airtight. But it is a concern,
and we are concerned about Iranian and WMD development.
Senator Cochran. I am going to ask you and also Secretary
Reinsch if you know whether any of the over 1,000 high-
performance computers exported from the United States since the
administration adopted this new policy have been shipped to
Dubai or anywhere else in the United Arab Emirates and whether
or not any of these computers has made its way to Iran.
Mr. Reinsch. Let me say first, Mr. Chairman, that if they
did make their way to Iran that would be a violation of U.S.
law. I would want there to be no doubt about that. And that
makes it for me an enforcement question, which has some of the
difficulties that both you and also Senator Durbin had
mentioned earlier.
We have a complete list of where they were sent. As I said,
in preparing for this hearing I focused on China and Russia. I
didn't bring the whole list with me. We can certainly find out.
It would not be a violation of U.S. law to export one to
Dubai, and I can easily find out if any were shipped there and
we have, as I said, means of determining whether they are still
there or not.
Senator Cochran. I am curious to know in your enforcement
activities whether or not you have undertaken to investigate
whether any of these supercomputers in the Middle East have
been trans-shipped to Iran.
Mr. Reinsch. We are looking at all of them right now. I
wouldn't say that we have an investigation of a specific one
under way. We are examining all of the records that we have
obtained and other information to determine whether there is
any evidence that warrants any investigation.
As Dr. Wallerstein and I both said beforehand thus far--
well, you've raised Iran. I guess our response was on China--
but thus far we have no evidence that anything like that has
happened. If we obtain any evidence then we will proceed with
an investigation. We are actively looking for it.
Senator Cochran. One question about the efficacy of these
new policies is the assumption that the administration seems to
be making that everybody can piece together small computers and
make these giant high-performance computers, 7,000 MTOPS and
higher, but I come back to this testimony that our General
Accounting Office gave in April that the United States and
Japan are the only countries in the world that can produce
high-performance computers operating faster than 3,500 MTOPS.
The White House, when it issued its fact sheet in October
of 1995 announcing this new policy said, in support of
President Clinton's statement, that ``we conservatively judge
that computers up to 7,000 million theoretical operations per
second will become widely available in open commerce within the
next 2 years.''
Is that borne out by the facts today or was that just flat
wrong?
Mr. Reinsch. I would say two things, Mr. Chairman.
First of all, we have another study underway to determine a
definitive answer to that question, which is why we have made
clear that we don't intend to take further liberalizing action
in the computer sector until a new study and further work is
completed, which may or may not recommend any further actions.
That study is not due to be completed until the end of the
year.
I would say, based on what we know and also looking at what
we are informed by industry in terms of technological advances,
that the 1995 study was somewhere between right on and
conservative in its predictions of what was going to happen.
And I would really recommend, Mr. Chairman--I don't want to
insert it in the record because that would kill several trees--
but I really recommend that you and/or your staff take a look
at the study. I think that it will make a persuasive case that
the issue in this area is not what GAO said it was in the sense
that it is not whether the Indians are building a machine. It
is not whether the Russians are building a machine. It is the
kind of processors, multiprocessors and work stations that are
out there via a whole range of producers that is the issue.
I think, in that regard, the study, if anything,
underestimated what has happened in the last 2 years. But as I
said, we will see. We are doing a new study, and we will be
guided not only by that, but partly by that when it is done.
Senator Cochran. I understand that you did base your
decision on a study, but that the study said that some of its
conclusions were based on ``conjecture'' and not hard evidence.
Is this the Goodman Study that you are referring to?
Mr. Reinsch. It's the study that I am referring to. I don't
recall the study saying it was based on conjecture, but I will
make a deal with you. If you will read the whole thing, I will
re-read the whole thing, and we can make a decision.
Let me say also that was, by no means, the sole basis on
which we did this. Each of the agencies involved, and Dr.
Wallerstein, I think, wants to comment, did their own internal
work on this subject.
Mr. Wallerstein. Yes. Let me add a few comments, if I may.
Senator Cochran. Please.
Mr. Wallerstein. First of all, with respect to the
assertion about the 3500 MTOP cut-off, you put your finger on
what the dramatic changes that are now underway. Up until the
1995 time period, we were able to measure the power of these
machines because they were so-called single vector processor
machines; that is, these were large boxes that had enormous
number-crunching capability. But now we are moving into a new
era, which is characterized both by massive parallel processing
and by clustered work stations.
So whole new strategies or architectures are evolving,
which change fundamentally the nature of the control problem.
This is, in part, what the new study will look at.
Our position has been--and this is a governmentwide
position since 1993--that we need to look at this approximately
every 18 months, not necessarily to change the policy, but at
least to determine where the technology and the markets have
gone in that time period, and that is what we have done and
will continue to do.
As Under Secretary Reinsch has already said, we are going
to look at it again. We may determine that the market has not
evolved that fast and, we need to look at where the
controllability thresholds are. And the report that he referred
to does talk about this notion of controllability thresholds.
So I do also commend it to your attention.
Part of what we did in 1995 was an internal DOD assessment
of the applications that we use computers for within our the
defense community. We discovered that there were clusters of
applications; one around 10,000 MTOPS, another above 20,000
MTOPS.
I might note, also, Mr. Chairman, that the most powerful
machines now are in excess of 100,000 MTOPS. So when we talk
about this range of 2,000 to 7,000 MTOPS, we are way down at
the lower end of a range that goes up to over 100,000 MTOPS.
Senator Cochran. And that is the market that I think you
were referring to, Secretary Reinsch, when you said we have
cornered the market on those high-performance computers. Is
that correct? We are the country that manufactures----
Mr. Reinsch. I certainly hope so, yes.
Mr. Wallerstein. These were highly sensitive----
Senator Cochran. What are your rules on that? Do we sell
those to anybody who has got the money to buy them or do we
have a list of prohibited purchasers?
Mr. Reinsch. Well, you have articulated the policy, Mr.
Chairman. With respect to Tier I countries, which are
essentially our NATO allies and a few others, we don't have a
limit. We do have a record keeping requirement. With respect to
Tier II, anything over 10,000 requires a license. I would just
say in passing, as evidence of what Dr. Wallerstein said, in
the first 3 months of this calendar year, we got more license
applications for computers above 10,000 than we had received in
all of 1996. And so this is a steadily and very quickly growing
field.
The processor data that I showed you, that I think I
mentioned in my testimony--I was going to blow these charts up,
but I think you can see them. This is the estimated performance
curve for single processors. Where we are on the curve right
now is this year. These are the industry projections for where
a single processor is going to be by the Year 2000. 2,615 MTOPS
estimated in the Year 2000. That is one processor. That is not
even a computer.
I have got similar data for multiprocessors.
And then, to me, the most interesting one is they did a
little upgrade chart; that is, a look at the extent to which
you could upgrade the existing box by adding new processors,
sticking in new boards. This year the range of upgrades that
you can undertake for computers, existing machines, ranges from
504 MTOPS to 122,000 MTOPS. I mean, this is mushrooming.
You may remember from hearings that I am sure you
participated in with respect to high technology Moore's Law,
which is computer speed doubles every 18 months to 2 years, and
this has been an axiom in the industry for I think about 15
years. I periodically ask industry people, ``Where is the end?
When do we cap? When can you not grow anymore?'' and they
continue to say never.
Now, I don't know whether I believe them, but every curve
we have seen from everybody suggests that that is the way this
industry is going very quickly, and it is going in exactly the
way Dr. Wallerstein said; clustered work stations massively
parallel processing. The big box is not what is happening any
more.
Senator Cochran. One question with respect to the fact that
your policies are based upon a definition of civilian use as
opposed to military use in determining whether or not a sale
would be permitted to a Tier III country. An export requires
our individual validated license if over 2,000 MTOPS, and can
be denied, if it's going to be for a military end user or for a
military purpose.
How can you tell in a country like China, for example,
where you have a mix of civilian and military activity at a
research lab like the Chinese Academy of Sciences, that the
supercomputer is not going to be used for military purposes in
some way or that any entity that is subject to influence by the
central government to share its technology with a military
entity is not going to be in a position to have to comply with
that?
It strikes me as very risky business, indeed, to permit the
sale of these highly sophisticated state-of-the-art
supercomputers to entities in China, which can easily pass on
the technology or share that with others in that country for
military purposes.
Are you satisfied that this is really serving our national
security interests? I am sure it is serving our economic
interests to permit these sales. But it seems to me that it is
putting our security interests at risk by carrying through with
this flawed policy.
Mr. Reinsch. I think you were right, Mr. Chairman, that in
China, not uniquely, but peculiarly, it is hard to tell the
military from the civilians, not because they hide it but
because the PLA has its fingers in a lot of civilian pies, as
it were, hotels, restaurants, things like that.
We rely on a lot of information. A lot of it is
intelligence based, and is based on information that we have
compiled through other enforcement activities over the years,
as well as the representations of the end users as well as
prelicense checks and a variety of other devices to make the
best judgment we can.
Now you have a colleague in the House, I know, who has
stated on the record in the hearing that we had on this subject
over there that, from his point of view, all Chinese end users
are bad because the PLA presumably is in a position to access
any computer shipped to anybody in China.
Well, of course, that is at least theoretically true in any
country in the world. As I said to Senator Durbin, once it
leaves our country, we lose control. Now, it is probably a bit
more realistic to suspect that that would actually happen in
China than it might happen in some other country. But if you
want zero risk, then you make a fair point about our policy.
And as I said in my testimony, the only way that you are going
to get zero risk in the computer business is to license
individually all PCs, including the ones in the back room, and
deny them all to virtually everywhere because even if you deny
them all to China you have got the secondary market and you
have got re-exports. If you are talking about thousands and
thousands of low-level computers, there is very little you can
do about it.
So we try to assess risk. We try to make our own
independent judgment of when an end user is a bad end user or
not. It is not our belief as an administration that all end
users in China are, by definition, bad and that all of them
are, by definition, under the thumb of the PLA, and we are
prepared to permit the exports of these things to legitimate
end users.
Now, I will also say that we have not approved any license
applications for high-performance computers to China of over
7,000 MTOPS. Under our policy, all computers over 7,000 require
a license. We have not approved any of those licenses. The
entire discussion that we have had today has been in the 2,000
to 7,000 range, where we have a distinction between military
and civilian.
You make a very good point. It is a very difficult judgment
to make and we and the companies are drawing the line as best
we can. I would never guarantee you that somebody isn't going
to end up on the wrong side of the line at some point.
Senator Cochran. I am reminded, when Secretary Perry was
here for his confirmation hearing in 1993, he was asked about
how you control the sale of dual-use technology, and he said,
``It is a hopeless task. It only interferes with a company's
ability to succeed internationally trying to control the
sales.''
Do you agree with that view, Secretary Wallerstein? He's no
longer the Secretary. You can disagree, you know. [Laughter.]
Mr. Wallerstein. I sense a trap here, Senator.
I have had the privilege of working with former Secretary
Perry on a number of studies before we both were in government,
at the U.S. National Academy of Sciences, on this very subject.
Secretary Perry is among the most thoughtful individuals on
this, both because of his defense expertise and because he is,
by training, an engineer and mathematician.
I do not know the context in which that question was asked
to him, but I do know that both in the studies that we
undertook at the National Academy of Sciences, and during his
tenure as Secretary, he supported carefully designed export
controls on dual-use technology. In fact, he was a strong
advocate for some of the policy changes that were undertaken
during the first Clinton administration.
Senator Cochran. Secretary Reinsch, have you turned over to
the Justice Department any evidence that would involve
officials of the Commerce Department in facilitating the sale
of these supercomputers to China that you think are illegal
transactions?
Mr. Reinsch. I don't know how to answer that question, Mr.
Chairman. Nobody has asked us for anything. The computers in
question, the computers that are under investigation were all
shipped without a license. There was no action by the Commerce
Department to permit those to occur.
I don't know what evidence there would be. If there is any,
I am happy to. Since we are not talking about licenses that
were approved, we don't--it is like trying to develop
documentary evidence on a negative. We don't have any
information. They didn't come to us.
Senator Cochran. Well, there have been some suggestions
that the Chinese Government has undertaken to try to influence
policies of this government by various means and through
contacts with various officials in our government, including
some who worked at the Commerce Department. And so I am curious
to know whether or not you have turned over the evidence of any
such transactions to the Department of Justice for their
review.
Mr. Reinsch. We've responded to every request and every
subpoena that we have gotten on all of these matters as a
department. The Bureau of Export Administration has gotten some
requests for information primarily from the Committees in the
Congress that are investigating the same issue. I don't recall
offhand if we have gotten a request from the Justice Department
or not, but we will certainly turn over to them whatever they
want, and we have turned over to the Congressional
investigators everything they have asked for.
The individual, I would just say in passing, the
individual, if you are referring to Mr. Huang, about who many
of the allegations have been made, was not part of the Bureau
of Export Administration and didn't interact with the licensing
process.
Senator Cochran. Senator Durbin.
Senator Durbin. No questions.
Senator Cochran. Thank you very much for participating in
the hearing and being here and sharing your testimony with us
and for your statements and for the additional material that
you may be able to give us to help us fully fill out our
record.
Thank you very much.
Senator Cochran. Let me introduce our second panel as the
Secretaries leave the witness table.
Dr. Stephen Bryen is the President of Delta Tech,
Incorporated. He has considerable experience in the field of
export control policy, having been the Deputy Under Secretary
of Defense for Trade Security Policy from 1981 to 1988.
While at the Department of Defense, Dr. Bryen served as the
first Director of the Defense Technology Security
Administration.
Dr. William Schneider also has experience in export control
policy, having served as Under Secretary of State for Security
Assistance, Science, and Technology, from 1982 to 1986. While
at the State Department, Dr. Schneider was the Chairman of the
Senior Interagency Group on the Transfer of Strategic
Technology. He currently serves as an advisor to the State
Department, as the Chairman of the Department's Defense Trade
Advisory Group.
Dr. Bryen, I am going to ask you to proceed first, if you
will. We have statements which we will put in the record, and
then we'll ask Dr. Schneider for his remarks, and then we'll
have an opportunity to discuss them with you.
Dr. Bryen, you may proceed.
TESTIMONY OF STEPHEN D. BRYEN, PRESIDENT, DELTA TECH
Mr. Bryen. Thank you, Mr. Chairman. I will submit my whole
statement for the record. I am going to touch on some of it by
way of introduction of this subject.
But before that I thought it might be useful just to
clarify a little bit what we are talking about, since there
seems to be some confusion between PCs and supercomputers in
the Department of Commerce, and since that confusion exists, I
thought I would try my very best to clarify, if I can.
Senator Cochran. Would you pull the microphone just a
little closer to you, so we can hear clearly?
Mr. Bryen. I will do the best I can.
Dr. Wallerstein mentioned two of three kinds of
supercomputers. There are three types known today and a fourth
that may emerge. There is the Vector processor, which is the
oldest type. The Cray computer is most famous as a Vector
processor.
There are massively parallel processors that are called MPP
computers, and then there is another type of parallel
processing called Symmetric Multiprocessors. So those are the
three kinds. PCs are not supercomputers and you can't stick
them together to make them into supercomputers, not yet. There
is work on clustered work stations, as Dr. Wallerstein
mentioned, but so far, at least, the breakthroughs have not
occurred and there is not that type of supercomputing available
to anyone yet.
I have no doubt that it will be eventually available.
The Symmetric Multiprocessor machine, which is the Silicon
Graphics type of machine that you have referred to, is one of
the most popular ones, and it is more and more used in the
Defense Department. And, in fact, in my testimony at the back I
have taken a look at just one supercomputing center in the
Defense Department. There aren't that many, but I took
advantage of the fact that this one was pretty well documented
on the Internet, and it became a convenient way for me to do my
research.
I might mention parenthetically that there is a lot of
information on the Internet these days, and it is one of the
good sources of learning about what the Chinese have been
buying in the way of supercomputers.
I would also like to make two ancillary points in respect
to that. The first is before the administration decontrolled
supercomputers--and this was about 18 months ago, I guess--the
sales were very tightly regulated. And as far as I know, not
one single supercomputer ever wound up in the wrong hands.
It's interesting that this new policy has caused a
diversion--I recognize Secretary Reinsch said that there are at
least two cases under active Justice Department investigation;
one in Russia and another in China. And that one in China is
the Chinese Academy of Sciences. So I have really very serious
doubts that this policy is serving our national interest if
this sort of thing goes on.
The U.S. Army site that I mentioned is very much like the
site that has been put into place in China at the National
Academy of Sciences. It has more or less the same machines. So
that is why I selected it. Other than the convenience of having
these nice graphics on the Internet is the fact that we can
take a look at what is going on there and the kind of work they
are doing.
By the way, I just was selective. There was too much to
fill up your book with, but if you would like to have all of it
I would be glad to print out the rest of it. Because some of
the work is in missiles, in-theatre missile defense, in
particular. Some of the work is in biological and chemical
defenses. Some of the work is in dealing with complex and
difficult problems such as how you design a hypersonic vehicle.
All of this is being done on this type of processor. This is
not child's play. This is a serious effort that the Army is
making to understand certain processes.
In one of the pictures you have is a re-entry vehicle. It
is modeled on the supercomputer. The idea is to make this re-
entry vehicle efficient and accurate so that it hits the target
that it is intended to hit with a high degree of accuracy and
that it doesn't fail in the process.
Another project going forward is studying how chemical
weapons disperse and how you can decontaminate, because we
faced that threat in the Gulf War, as you know, and, as
everyone recognizes, we are going to face it again. So
understanding how to deal with it is very important.
While they are looking at that, the Army is also looking at
how you clean up a subway station. You know what happened in
Japan. It could happen here.
This is the sort of technology that we are selling to China
and to other countries.
I also want to say that you can't hook up a lot of PCs and
get a supercomputer. I don't know where that idea came about.
If you could do it, the Chinese would be more than happy to
hook up a lot of PCs and not buy million-dollar supercomputers
from Silicon Graphics or Hewlett-Packard or Digital Equipment
or any other company.
The fact is, when the flood gates were open, they bought a
huge number of supercomputers. Most companies that sell these
things--they don't sell that many of them--most companies that
sell these things are pretty proud of the sales, and they
usually post them in their publicity. They put out a press
release, ``We just sold a supercomputer to this company or that
company. Isn't it great?''
Silicon Graphics did that, and you can get on the Internet
all of their sales up to the time of the decontrol when they
stopped publishing the list. I think the answer is clear. Their
biggest customer is probably China now, and they are not real
proud of that or at least they don't want us to know too much
about what's going on.
But what's going on is that a lot of technology is being
transferred to China. I am very concerned about it. I think it
is not in the national interest. I think that these machines
can be controlled. I think that we could have a policy that is
effective.
The other thing I would like to point out is that these
machines are not standalone. They are not just being sold to an
end user. This is a myth. It is total mythology. They are being
sold and put into networks, hooked up to all kinds of
institutes. Defense establishments I am sure are connected.
Nuclear establishments I know are connected. Universities are
connected.
More than that, these networks--by the way, the networking
technology is also coming from the United States. From what we
can determine, and this is from outside research, the networks
are both public and private; that is to say, there is a
classified portion to the networks that are being established.
So that what you see is not what you get. What you see is the
public part, but there is a private part, classified part,
probably where most of the sensitive nuclear and other kinds of
research is going on.
We know that China is pushing very hard to modernize its
military forces and its nuclear weapons. I will just refer to
two areas, but Secretary Albright also mentioned a third one
yesterday, a new longer range ICBM that China is working on.
But in addition to that, China is working on a MIRV capability,
a Multiple Independently Targeted Re-entry Vehicle capability,
and the ability to build small and compact nuclear weapons that
can work in a MIRV missile.
Supercomputers will speed up the process of all of that and
make it, I believe, possible for China to achieve that kind of
capability very quickly.
In addition to that, China is working very hard on cruise
missiles, and I am not completely sure in my own mind whether
cruise missiles aren't worse in some ways than ICBMs. That is
because you can't tell what kind of warhead they have. They
could have a nuclear warhead. They could have a chemical or
biological warhead or a conventional warhead. We used our
Tomahawk cruise missile in the Gulf War, as you know, and after
that, most recently, in a retaliatory attack with conventional
warheads. But that same missile can carry a nuclear warhead.
And China can design the warheads and, in fact, a lot of
the whole vehicle using supercomputers.
In addition to that, China is acquiring other technology
from the United States, which is being approved by license by
this administration, such as the Global Positioning System
manufacturing technology, which enables China to have high-
class guidance at a relatively low investment and to do it
quickly.
I think the issue here is the speed at which China can
modernize its military capabilities, and what we are doing is
aiding and abetting the process of giving to China, selling to
China, if you would like, these sorts of capabilities.
I think that this policy on supercomputers is a very
dangerous one. You have had a chance to talk to the Secretary
about that. He claims that the Commerce Department is
intimately involved in all of these decisions, even though
there are no licenses here that they consult with all of the
companies.
I can't, for the life of me, understand how, for example,
they would sell a supercomputer system to the Chinese Academy
of Sciences. I have been there. I have seen their nuclear
accelerator. They showed it to me. I know what kind of business
they are in, and I think the administration knows what kind of
business they are in.
Why we would sell supercomputing to them is hard to
understand. It doesn't really make sense. And, again, their
system is part of other networks. All of this is tied together
by fiber optic high-speed networks, again sold by the United
States.
So it would seem to me that what really should be done is a
pause, a halt. Such exports should be stopped for now. I have,
at the end of my testimony, a number of recommendations. They
are not my recommendations. There is an organization here in
town called the Jewish Institute for National Security Affairs,
which has, among its members, many retired military and flag
rank officers who look at these questions, and their sense is,
first, to suspend the current regulations on high-performance
computers and require individual licenses for them. I am not
saying you can't export them, but get a license like we have
always done in the past. There aren't that many, 47 or whatever
the number is not going to exactly strain the capabilities of
the Defense Department or the Commerce Department to process
licenses and to have real accountability on these transactions.
Secondly, let's get a full accounting of what is gone. I am
far from convinced that 46--or now it is up one more
officially--47, is the right number. I think it is a much
larger number, but I can't say for sure, and I think that we
are all owed an explanation as to what's actually transacted.
I know Commerce Department loves to wave the flag of the
12(c) Export Administration Act provision. We can't tell you
about this. It is proprietary and all of that nonsense. Then
just don't put the company names. Just tell us where they went,
how many there are, what the speed of the processor, and then
they are completely out from under 12(c). 12(c) is only
designed to protect corporate proprietary information. So if
you take away the name of the company there is nothing to
protect, and they can provide that list today or at least what
they have, and the Committee should have that, and the American
people should have that, should know what's going on so that
independent people can make an evaluation, since, quite
frankly, the administration has not made any.
There is no study that I know of that has really looked at
the military implications of any of this and, specifically, in
relation to China. I don't know of any. I heard a lot of babble
about that in the conversation today, but none was really
referred to, and I don't think there is one, and that is very
scary, very scary, and there should be one.
Third, we should have a study of the impact of computer
sales both on our national security and on weapons
proliferation. The proliferation issue is a very serious issue.
China is a proliferant country. They have been selling missile
technology to Pakistan. That has been the subject of quite a
lot of controversy in the past few years. This is a serious
matter. It has lots of consequences and surely we should try to
understand what is going on and do so soon.
Fourth, using the CIA and the Defense Intelligence Agency,
and I am more partial to DIA in this, let's see who is trying
to get supercomputers and what their reasons are, what do they
want them for.
Everyone talks about we're doing civilian research. What is
that? I mean, what is really being--what are they really being
used for? What kind of work are they being used for? Do we have
any idea?
I think the intelligence community can tell us a lot about
that if they put their mind to the task, and they surely should
be tasked to do that soon.
And, finally, and most importantly, develop and propose an
effective multilateral export control licensing system, not one
that fails on the most crucial issues.
Again, I reiterate. We never lost a supercomputer before.
This administration has lost at least two in Russia and one in
China, which they admit to, and probably a lot more, and that's
something that we should seek to prevent, and I believe that we
can do that. But it is going to require cooperation.
As far as who owns this business, Senator, you are both
quite right. It is really the United States. Japan has
supercomputers, mostly the Vector type, but the real parallel
processing type machines are being built here. So it is our
technology, and I think we have the possibility of controlling
it.
I should also add to that, that it's not just the hardware
that is at issue. The software is very important because one of
the things that we learned with respect to what the Russians
were trying to do during the period of the Soviet Union is they
were trying to get Western hardware, but they were also seeking
to run Western software on it.
A lot of development work in software is very critical to
how you build weapon systems and, consequently, I think you
have to control the software as much as the hardware in some of
these cases.
So that's my way of introduction, and I will be glad to
answer your specific questions.
Senator Cochran. Thank you very much, Dr. Bryen.
[The prepared statement of Mr. Bryen follows:]
PREPARED STATEMENT OF DR. STEPHEN D. BRYEN
Senator Cochran. Dr. Schneider, welcome. You may proceed
with your remarks.
TESTIMONY OF WILLIAM SCHNEIDER, JR., FELLOW, HUDSON INSTITUTE
Mr. Schneider. Thank you very much, Mr. Chairman. I
appreciate the privilege of being able to appear before the
Subcommittee today. I have prepared testimony, which I have
already provided, and I will simply offer a few observations
and also some suggestions for proposed reforms.
The administration has appropriately placed the
counterproliferation struggle at the top of its foreign policy
priority agenda. The problem I have is that the export control
system, in parallel with diplomacy, are the front lines of
dealing with the proliferation problem and are not likely to be
effective in achieving the laudable counterproliferation aims
of the administration. This is so not only with respect to
weapons of mass destruction, but also to their means of
delivery and advanced conventional weapons as well.
We know the threat of proliferation is a very serious one.
Just to give the Committee a calibration point about how
extensive the sweep of the decontrol on sophisticated
technology is, is during my own service in the Department of
State a decade ago, in the area of dual-use export licenses,
approximately 150,000 licenses were issued each year during
that period.
Now although the volume of trade has more than doubled
since that period and the volume of high-tech trade has gone up
several fold beyond that, we are issuing only 8,000 licenses
per year. In other words, from 150,000 down to 8,000. It is a
very substantial scope of decontrol.
A second point is it is important to take note of the
changes in the way military technology is developing. Until the
1970s, military technology was fairly isolated. It was
developed uniquely for military applications and it typically
was at the cutting edge of the employment of modern technology.
The situation has changed very rapidly with advances in
microelectronics computation, advanced materials, and so forth.
Now the engine driving military performance is sophisticated
civil sector technology. This is creating what I have described
as a ``new path to proliferation'' compared to what we had
anticipated a decade ago.
Rather than seeking to acquire information about the
scientific trick, so to speak, to, say, produce nuclear
weapons, which we have historically guarded very carefully by a
statutory classification system. Now proliferation is a
question of industrial processes. The scientific knowledge or
secrecy surrounding the scientific knowledge no longer protects
us from the proliferation of these advanced technologies. The
proliferation of industrial processes is what is creating the
problem.
Two recent examples: Iraq and China. In both cases, we have
had a policy of not providing them with munitions list
technology; that is, defense-related technology, but we did
have, until the Gulf War, we did have a policy of allowing Iraq
to have access to sophisticated civil technology.
The result was that they were able to produce not only
weapons of mass destruction or move considerably towards that
goal, but also ballistic and cruise missiles.
The case is even stronger with respect to China, where the
United States and most of the European allies do not ship
advanced munitions list technology. The access to sophisticated
industrial technology, including computers, machine tools,
software materials, et cetera, is accelerating the rate at
which China is able to modernize its Armed Forces. Facilitating
the modernization of China's Armed Forces is not U.S. public
policy and, as a consequence, the aims of U.S. policy are being
frustrated by the ineffectiveness of the export control system.
As a consequence, I have suggested a few reforms that might
be considered by the Congress and, perhaps, by the
administration as well. One is to refocus the policy on
controlling exports to get at the proliferation problem.
Decontrol has gone to the point where the export control system
is, in effect, abetting proliferation rather than containing
it.
The second is intelligence collection and processing.
Intelligence collection and processing is essential to
effective diplomacy in the counterproliferation field. The
level of effort and the nature of intelligence support to the
export control function has declined very substantially, and in
order to reclaim the effectiveness of the export control system
this needs to be changed.
The third point is an export control regulatory practice.
The U.S. Government should be maintaining a data base on end
users. The problem that Dr. Bryen mentioned in China is a
serious one. The PLA owns more than 20,000 businesses and many
of them are fronts for defense-related transactions, and this
needs to be monitored.
Restoration of end user checks in China is another issue
that should also be considered. I think this can be done
without significant changes in appropriated funds. As I
mentioned, the Bureau of Export Administration and the
Department of Commerce has about 300 full-time equivalent
personnel to process about 8,000 licenses.
The Department of State has munitions list licensing
responsibility. They have 45,000 licenses, and they have less
than 50 people processing the licenses. So I think there is
enough in the way of head count in the Agency to support more
end user checks to these sensitive destinations.
The fourth point is to increase diplomatic support for
export control management. Dr. Bryen mentioned an effective
multilateral regime. I think a dimension of this is effective
diplomatic support to work with other countries that are
becoming diversion channels for some of this technology or are
in other ways abetting the diffusion of the technology.
The final point is interagency coordination. We found
during the COCOM period where export controls were a high
national priority that it was important to have effective
senior-level policy coordination to make sure that all of the
agencies were using their respective resources to bring the
matter to closure. The decline in the importance of export
control is reflected in the diminishing level of bureaucratic
attention it's receiving, and I think the interagency process
should be effectively reformed.
I am pleased to entertain any questions you might have, Mr.
Chairman.
[The prepared statement of Mr. Schneider follows:]
PREPARED STATEMENT OF WILLIAM SCHNEIDER, JR.
Senator Cochran. I appreciate very much both of you
touching on not only your impressions of the comments that were
made by the administration witnesses on this subject, but
making some suggestions for reform and change in the policies
and initiatives that have to be undertaken if we are going to
protect our security interests in connection with the export of
these dual-use technologies.
There was one suggestion that we heard that in the case of
Russia one of the computers that they ended up with there ended
up coming from a U.S. company in Europe. What can we do about
that, the trans-shipment? I asked the witnesses about exports
into the Middle East that could wind up in Iran. Of course they
could end up in Iran from China, too, because not only has
China sold missiles and technology to Pakistan, but it's also
sold cruise missiles and other technologies to Iran as well.
How do we better protect ourselves from the trans-shipment
of supercomputers? Is there any way to do that?
Mr. Schneider. Yes, there is, Mr. Chairman, and this is
quite readily attended to if the product is licensed. This
requires effective intelligence support to be aware of
diversions if it is indeed a criminal diversion where the, say,
a buyer in Europe is trans-shipping it to a bad end user, let's
say, in Russia. If we know about it, the customs protocols are
in place to enable us to stop that shipment, and this can be
done, again, with effective diplomacy.
But as Dr. Bryen was suggesting, we have to have a
licensing regime for these, otherwise it's not practical to
effect those controls.
Senator Cochran. Dr. Bryen.
Mr. Bryen. Well, I agree completely that the whole issue
here is the administration is allowing companies to send these
computers out of the United States without licenses--without
licenses--and then you can't enforce anything. It is impossible
to account.
Even the accountings they get are only first order; that is
to say, where they think they went the first time. Where they
end up, they have no idea. So you have to have individual
validated licenses if you want to have any accountability. It
is that simple.
Senator Cochran: What suggestions with respect to the
definition of high computer technology? We were talking about
these tiers and the tier that is described as being 2,000 MTOPS
to 7,000 MTOPS. Is that relevant to anything? Is there any
basis for the tier system and the definitions in terms of MTOP
capacity?
Mr. Bryen. I think the whole thing is totally synthetic,
first of all. The three tiers or actually they call Tier IV--
one is a ``no'' tier--but the three ``yes'' tiers consist of a
grab bag of countries that are very peculiar, lumped together,
and it doesn't make any sense. You have lots of little
inconsequential countries in the second tier, for example. They
can get very powerful machines. What they would use them for no
one has any idea.
So this business of tiers, I think, is bizarre.
Secondly, the number, this 2,000 threshold number has been
pulled out of someone's hat. I don't think it has any relevance
to anything. You have to try and grasp, basically, two things;
one, how machines would be used, what purposes will they be put
to, and try to peg your controls on that information. If a 500
MTOPS machine is what you need to do good, solid nuclear work,
then you should be controlling those, not 2,000 MTOPS machines.
And I don't know that the Goodman Study or any other study has
actually addressed that important issue.
So the first point is to try and figure out the end use and
then try to peg your computer controls to that.
The second issue is the fact that parallel machines,
particularly the symmetric ones, can be hooked together, and
they become more powerful machines. This is a considerable
concern. If you can sell four or six or eight or 10 or 20
machines that can be stuck together and it would be something
else, then you have a real problem on your hands.
The current regulations don't distinguish the type of
machine; that is to say, they don't deal with whether it is a
parallel machine or a Vector machine or what kind of machine it
is, whether it can be effectively coupled with another sister
processor to make a more powerful unit, and I think the
regulations there need changing. They need to reflect the real
power--MTOPS is not enough. You have to have more about the way
these machines work together, the different processors work
together in order to make a judgment on how to handle this.
Senator Cochran. Dr. Schneider.
Mr. Schneider. Just a footnote to that is that the
technology of networking the computers, in addition to the
computers themselves and the associated hardware, should also
be the subject of the control regime because they are
increasingly crucial to the exploitation of these capabilities.
Senator Cochran. There was a question I asked of the
administration witnesses about the evaluation or assessment
that had been done by the military of the potential threats to
our national security by reason of exporting supercomputers. I
am curious to know what your information is about whether or
not we have undertaken, to your knowledge, any such assessment
of the threat to our national security from the sale of
supercomputers.
Do you know of any that has been done by our military
officials?
Mr. Schneider. No. The one study I am acquainted with,
which was, I believe, done in response to a House National
Security Committee request and legislation a year or so ago was
done by a group at Stanford University. It was not done by the
Defense Department itself. It was apparently contracted for
study, but I don't know of one that has been done by the
Defense Department itself.
Senator Cochran. But this is a recent study that was done
for the Department of Defense; is that correct?
Mr. Schneider. Yes. It was done for the Department of
Defense, but it had some limitations in the way the study was
done, so it would not be able to respond to the questions you
were asking.
Senator Cochran. Dr. Bryen.
Mr. Bryen. I don't know of any Defense Department study
that relates our national security to the sale of
supercomputers. I don't think there is one.
There was an Energy Department study in the late 1980s,
which assessed the importance of supercomputing to nuclear
weapons development. That's a good study. The Committee ought
to get one. It's not classified, and it makes some very telling
points about how supercomputers are vital to the design and
deployment of nuclear systems.
Senator Cochran. Based on what you know about our policies
and about the emerging sophistication of weapons of mass
destruction in both China and Russia, particularly, do you feel
this is a situation where we ought to press the administration
to make a change in its policy? Are we on the right track here
in this hearing by pushing for this and urging that it be
reconsidered?
Mr. Schneider. I believe so, Mr. Chairman. I think there is
a wide consensus, not only within the Executive Branch, but
between the Executive Branch and the Congress, about the aim of
controlling proliferation. Export controls are a crucial aspect
of it, and it is the part of the system that is broken.
Senator Cochran. Dr. Bryen.
Mr. Bryen. I agree with that. I should also mention the
administration is trying to decontrol even more supercomputers,
and that is what the second study is all about that they are
conducting at the moment. The word is, that I have heard, is
that they want to set the base threshold at 10,000 MTOPS
instead of 2,000 MTOPS, and they are trying to justify that. So
anything you can do, Mr. Chairman to block progress in
decontrolling even more sensitive technology would be very
helpful at this stage.
I am very worried about this. This is I would call it
reckless, sir.
Senator Cochran. Well, I appreciate very much your
attendance at this hearing and your very helpful participation.
I am going to call on the General Accounting Office and
request an additional study of the relationship between the
dual use export control liberalization of this administration
and our national security. They have had an opportunity, and I
referred to a report that they had done already on this subject
of export controls and supercomputers. But I think we need more
information, and since we don't have it from existing sources,
we'll ask GAO to do one.
I think this has been a most interesting and helpful
hearing. It has also been disconcerting and alarming in many
ways.
We are going to continue to pursue our interest in
proliferation and our national security interests. Our next
hearing is going to be on July 17 at 2 p.m., where we will
review the ABM Treaty Compliance Review Process. Until then,
the Committee is in recess.
[Whereupon, at 11:55 p.m., the Subcommittee adjourned,
subject to the call of the Chair.]
(all)
|
NEWSLETTER
|
|
Join the GlobalSecurity.org mailing list
|
|
|
