Room SD-106
Dirksen Senate Office Building
Washington, D.C.
Friday, January 19, 1996

CHAIRMAN BROWN: I suggest that we begin as nearly on time as we can. We have a full afternoon schedule, especially if all the witnesses manage to make it here through the weather. So I will call the Commission hearings to order once more, and thank our next witness, Dean Joe Nye of the Kennedy School of Government at Harvard.

Dr. Nye has served his country in the Defense Department as Assistant Secretary of Defense for International Security Affairs; he's served in the Intelligence Community as Chairman of the National Intelligence Council, in fact, that was his most recent position in the government; and in an earlier administration he was Deputy Under Secretary of State. So he has varied government experience in several different departments and agencies, and of course has a very distinguished academic and intellectual record, having served on the editorial boards of several periodicals that deal with foreign policy issues, and is a well-known author, both of articles and books.

Again, we welcome his appearance, and we look forward to his comments from that broad background of experience as a theorist and as a practitioner.

Please go ahead, Dean Nye.


MR. NYE: Thank you, Mr. Chairman. I will try to keep my opening remarks quite brief, because in the hour we have I think it's more important for me to answer your questions than for me to tell you my ideas as I see them. But I do have a very brief opening set of remarks.

I think if we look at the problems of intelligence analysis, where I will focus my comments, we find that it's become much more difficult with the end of the Cold War, for several reasons. One is the change in the structure of world politics has made a much more complex world. In a bipolar world where you had the U.S. versus the Soviet Union your intelligence priorities flowed quite naturally out of the structure of power, the structure of the world balanced power.

I have written, and won't go through it all in detail, that what we have today is something like a three dimensional chess board in the structure of power internationally, where the top level or the top military board is unipolar, with Americans the only great military power.

The middle level, the economic board, is multipolar, with the U.S., Japan, Europe, and China coming up fast in the sense of four major powers.

And the bottom level of the board of transnational relations, of things that cross borders from the maligned into the spectrum like money flows to the -- or the benign end of the spectrum like money flows to the maligned end, like drugs and terrorists. In this area there is really no distribution of power. It's totally fragmented, and almost nobody is in control. The net result of that is that the analyst has a much harder job deriving priorities from just looking at the international balance of power than was true during the Cold War.

The second problem is the diversity of the issues, which flows from the description I just gave of what's happening in world politics. It's often very hard to know which issues we should be focusing on, and some things that look like obvious areas to cut back then suddenly turn out to hurt us after we've made the cuts. After the end of the Cold War it looked pretty clear that Somalia and the Horn of Africa was no longer relevant to the world balance of power. But soon after we began to cut back resources devoted to Somalia we found ourselves involved with Somalia and desperately needing intelligence on Somalia because of a total different type of analysis which didn't start with balance of power politics, but started out of the role of TV and humanitarian crises.

A third problem we have is the problem of trying to relate politics and technology. Sometimes we can make a projection that's fairly clear in technological terms. This happens very often in the nuclear nonproliferation area where you can say does a country have certain facilities, how long would it take to construct those facilities, what's the range that it would take them to make a nuclear weapon?

For example, one could say that Iran would take about 10 years for them to make a nuclear weapon given their indigenous capabilities. But if you suddenly put in politics, such as the changes that are accompanying the disintegration of the of the Soviet Union and the fact that now you have a great deal of nuclear material and technological expertise that could be suddenly imported into Iran by various black markets and transfer of people, you might have to revise those estimates very dramatically. In other words, the linear projections can suddenly take a sharp knee in the curve because of the unexpected imports that come from political changes.

Yet a fourth reason why analysis is more difficult is what I've sometimes called before and others have called the difference between mysteries and secrets. Secrets are things which you can steal, like the size of a warhead on the SS-20. Mysteries are things which it doesn't do you any good to steal, because the people you're stealing from don't know the answer, either, such as will Yeltsin be in power a year from now? Yeltsin doesn't know the answer to that.

The Cold War, I think, was a setting in which many more secrets were important. And in the period after the Cold War I think a higher proportion of the questions that decisionmakers and policymakers want answered are mysteries. And that leads to a problem in terms of how you do good analysis. If you use your clandestined means and focus too heavily on just stealing secrets, and you think that that's where the value-added comes from, you may miss the fact that a great deal of what's good analysis of mysteries comes from diverse, open sources.

In that sense, the problem that we have, to use a metaphor, is that sometimes the intelligence community has felt that as it assembled a jigsaw puzzle that it had several very nifty pieces which had been obtained by clandestined means, but it often had difficulty seeing the picture of the puzzle as a whole that was on the cover of the box. And it was those open sources, those outside sources, which often provided that information or that perspective that allowed you to determine where the clandestinely obtained piece of the jigsaw puzzle really fit.

And finally, I would say that the revolution in military affairs, in which we are looking at a great deal of demand for support of intelligence for a military that relies upon precision operations, has greatly increased the cost, the needs and the costs, in terms of support for the military. So these add up to an environment for intelligence and intelligence analysis which is much more complicated than we've had in the past.

Let me say a word now about how intelligence reaches policymakers, which I can testify to from having sat on both sides of the table, so to speak, including quite recently. What I think intelligence producers sometimes fail to understand is the perspective of the policymaker in terms of time. When I was in my office in the Pentagon, the idea that I would have time to sit down and leisurely read through analyses was just out of the question.

What would happen is that I was on a constant whirlwind of activity, and the way intelligence came to me, the most useful way, was either in a very short written summary or in an oral briefing. So in the morning the Defense Intelligence Agency would come in and have for me an intelligence book which included all the different agency's production, but with a 2-page summary at the front which I could glance at and get at least the headlines in 5 minutes. And sometimes it was even difficult to get those 5 minutes before the phone was ringing or I was due at a staff meeting or what I thought was a half an hour to read intelligence was interrupted by a call from the Secretary or a Senator or what-have-you.

So that short here are the headlines for current intelligence was extremely important, and very often I would try to read the rest of the intelligence before I left the office in the evening. Sometime at 7:30 or 8:00 I would pick up the book and then try to read through the NID or the INR report and the other pieces that were in the book. That was often cut short by the fact that I had to go to an Embassy for some dinner or by the fact that my wife would complain when I never showed up at home for dinner. So in that sense, one of the great problems of the intelligence analyst is to get a feeling for how the policymaker is constrained in terms of receiving intelligence.

What I find particularly useful was to distinguish between two types of intelligence, current intelligence, of the type I just described, and estimated intelligence, the type of intelligence that gives you insights into what's the larger perspective of what's going on. In current intelligence it was important for me to have it in very brief written form or oral form so that I could get the headlines I needed to start the day and get them quickly.

Estimative intelligence worked in a different way. What I would do is I had a member of my staff who I would tell to stay in contact with the work that was being done in various intelligence agencies, and particularly the national intelligence estimates, and say when you see that there is an issue that we need to step back from the in box and get ahead of the curve on, schedule a meeting. Go to my scheduler and put it on the calendar. Put down a half an hour or an hour we're going to look at what's happening on North Korean economy, or we're going to look at what's going on in the underlying changes in the Persian Gulf area, or what have you.

And then we would collect a group of policy people, bring in the National Intelligence Officer, the Defense Intelligence Officer, and others, and structure a conversation in which they would start by telling us what they thought was happening, often based on a recent national intelligence estimate. But by staying in the room and being there as opposed to just reading it, we could press them. We could say, well, you haven't told me enough about that, or don't go on about that, we understand that. And that oral interchange was extraordinarily useful.

It also was extraordinarily useful in the sense that we would use each of these sessions to bring it to a policy point at the end of the session. In other words, having heard the intelligence people, we'd say we think we ought to be doing this, here's the direction we ought to be going. And we wouldn't ask the intelligence analysts to tell us what policy should be, but we often would say to them if we go in this direction, are we going to go over a cliff? Is this going to work? Can you caution us about whether this is going to be a disaster or not? And in that sense, that interaction between intelligence and policy was extremely effective.

What this suggests to me is that intelligence needs to be presented in a brief format, but we should be careful to distinguish between the current intelligence and the estimative intelligence. If you rely only on current intelligence in its brief form it's like reading only the headlines in the newspaper, and never reading deeper.

An estimative intelligence is extremely important, but it's going to be essentially demand driven. It's going to be something when I feel a need or the policymaker feels a need, then we want to draw on these people. It's not something where you're going to be able to have the time or the luxury to read it on a regular schedule. This means that we have to address the question of how do intelligence analysts find out what's the right timing. When is this going to be the front-burner issue? And the ways in which to solve this, I think, are to have a lot of contact between intelligence analysts and their policy customers.

For example, in the morning staff meetings which I held with nine deputies in the policy area that I supervised, I always had the intelligence analysts from the Defense Intelligence Agency sit in, so that she know what was on our collective minds. And very often, she would be able to make sure that something was brought to us later in the morning or we could ask her questions, is there anything new on this? So by being in the staff meeting, she knew what was happening, and that gave a great advantage to us, and also to her.

Similarly, on these estimative intelligence issues, by bringing in the people from the National Intelligence Council, the National Intelligence officers, and having them sit there with us, we were not only able to ask them questions that went into greater depth than areas where they may not have provided all the details we wanted in the national estimates, but they also learned much more of what was on our mind. And one of the things I think is really critical is that we not rely too heavily on a written product. I sometimes call it Nye's paradox of government that government is a machine designed to produce vast amounts of paper, and yet the higher you go in the government the less paper matters and the more the critical communications are oral.

I think the lesson of that is that you have to have interaction between intelligence analysts and policymakers. And I've suggested a couple of ways to do that. Another way is to have more intelligence analysts spend time in policy shops, in rotational assignments. In fact, I did this when I was in the Defense Department with two of my deputies I brought in from the National Intelligence Council. They turned out to be extremely good in the job they were doing. I think they will be even better intelligence analysts when they go back to the intelligence community.

I also believe that rotations of this sort, getting people around and about different agencies, is good government in general. It's a good way for people to understand what's going on. One of the greatest problems are horizontal communications in government. So I think it's not only a case in point but a broader good management principle.

Now, sometimes people will say, when they hear me describe this, well then why do we need a CIA? Why shouldn't all intelligence be devolved to the policy agencies, INR for the State Department, DIA for Defense, and abolish CIA? I think the answer to that is that you want to have multiple sources of analysis, particularly when you're dealing with mysteries rather than secrets. After all, the analytical part of the overall intelligence budget is less than 10 percent. We can afford to have a healthy duplication in analysis.

It's often predictable that INR analysts will sometimes come out more optimistic on a subject, that DIA analysts will come out more pessimistic, and CIA will be in the middle. This isn't always true, but I'd say in eight cases out of 10 it probably is true. And it's not because people are corrupted by policy interaction. It's quite natural. If the intelligence analysts is from State Department, always talking of the diplomat whose job is to pick herself up off the floor and go into the game again and try for another diplomatic solution even though the odds are 1 in 100, you're always looking for that little glimmer of hope.

And if, on the other hand, you're the intelligence analyst who is going to talk to the general who has to say don't give me your best estimate. I want information about your worst estimate, because if the worst comes true I've got to take forces there, and I want to know what I'm going to face, then you're going to shade things on that side. So I think multiple sources in the form of having a CIA as well as a DIA and an INR, and then having a way to pull them together and to relate them to outside sources of information, is the best way to structure intelligence analysis.

I have written and spoken elsewhere of the fact that I think a division of labor is to keep the three analysis groups that we now have, but to have a distinction between current intelligence and estimative intelligence based along the lines of the major agencies doing current intelligence and a separate National Intelligence Council, indeed it might be renames a national estimates council and brought out of the location in Langley to some place in F Street, which is what's planned, as an area where you could actually bring about this coordination and the estimative side, while still keeping that a very close sensitivity to policymaker's needs in the current intelligence area.

So these are some opening thoughts for you about the problems of analysis, why it's more difficult after the Cold War, the experiences I saw as sombody who worked both in intelligence and policy of what made intelligence useful from a policymaker, and a rough approximation of what I would do in terms of structure, in terms of making this more useful for policy.

Thank you.

CHAIRMAN BROWN: Thank you very much, Dr. Nye.

At the beginning of your statement you pointed out that there is, after the end of the Cold War, a variety of situations, locations, incidents that are potentially injurious to U.S. interests. And that means that parts of the world and groups that were of no great interest before, are now. We have not fostered the same degree of language skill, political knowledge, and so forth, about these things that we understandably concentrated on the Soviet Union and its allies and clients during the Cold War. But there is a lot of information about such diffuse and widespread possible threats among academics, business people, others in the private, and there is some, obviously, among diplomats and other government officials.

A lot of that expertise, a lot of that information, is not secret, and therefore, an important function, more important than it was before, is to synthesize, Well, acquire, and synthesize, that sort of knowledge, most of it nonsecret, add judgments, and then use those judgments to solve mysteries rather than to unlock secrets. Some of the things you want to find out about are short-term; some are long-term. My question, then, which you have addressed in part in your last remarks, is "How is the Government to carry out that function?" Begin with the front end, and take it through the process. Supposing that you are trying to find out something, you're trying to solve a mystery where the input is almost all unclassified, but it's not resident in the government.

Who should do that? Where do you locate it? Do you put it in the NIC? Do you physically move it, as you have suggested? Do you put it in a new agency? Where would it report if it were in a new agency? Or could you contract out for some piece of it, at least, the front-end piece of it? And if you contract it out, where would you contract it out from in the government?

MR. NYE: I think the challenge for current intelligence is to ask what's our value-added that CNN doesn't do? And the challenge for estimative intelligence is what's our value added that the economist or the financial times doesn't do? Just to be impartial among American periodicals. And those are hard tests, because a lot of the answer to mysteries is in first-rate journalism.

CHAIRMAN BROWN: Some of it must be in academia.

MR. NYE: And some of it's in academia.

CHAIRMAN BROWN: More, now that you're there, of course.

MR. NYE: But then if you ask what, then, is the unique contribution of intelligence on estimative intelligence, it's the fact that the policymaker just doesn't have the time to read those academic articles, or even to read the economist in the Financial Times in a timely way. I was usually 2 or 3 weeks late when I read the economist, and that was on weekends and at the cost of other things. So one thing that the intelligence analysis can do, estimated intelligence analysis, is distill this information in ways that are relevant to the policymaker, bring it to the point that the policymaker needs.

The other thing, obviously, that the intelligence analyst can do, doing estimative intelligence now, is to add the clandestined nuggets, which can make a difference. Suppose 80 percent of the solution to a mystery is in the open source, there being another 20 percent that might be clandestined, and that will vary by topic. If you're dealing with something that's related to North Korea it may be 90 percent that's clandestined and 10 percent open source. If you're dealing with something on the environment or AIDS or economic intelligence, it may be 90 percent open source and only 10 percent clandestined. But the combination of adding the clandestined and the distillation for policy relevance is something that I think needs to be done if you're going to provide good estimative intelligence.

I would like to see the National Intelligence Council play that role. I think it's properly structured, it's small, it's flexible, it brings in people from outside, it brings in people from other agencies, and I think one of the problems for the NIC in the past is that it's been seen too much as an adjunct of CIA. It's located just down the hall from the Director of Intelligence. That at least gives it access to the DCI, but it does mean that the other agencies don't see it as as much an impartial coordinating body as it might.

In addition, I think that this body that does estimative intelligence ought to draw on more agencies of the Government. When you're talking about financial matters, Treasury has expertise that should be brought in. Trade matters should be done from Commerce. So I would imagine having a -- let's call it a National Estimates Council or a National Assessments Council that would be located on F Street or in the Old Executive Office Building, somewhere in that area, which would draw on outsiders from academia, draw on members of other Government agencies, so among the ones I just mentioned, and draw on the various intelligence agencies which would produce a unique product which would be tailored particularly for policymakers.

Now, should that be under the control of the DCI, or should it report to the National Security Advisor? One could imagine either model being satisfactory. If you believe that you want to make sure that you don't weaken the DCI, I would tend toward leaving it under the control of the DCI. I think that would be my preference. But I could also imagine a different model of how it might report directly to the National Security Advisor.

I suspect the National Security Advisor doesn't have the time to actually manage this in the way that the DCI would, so that's why I would prefer it that way. But that's a model then of handling this puzzle, which I think answers the question we have to start with: How can we add something that you can't just get out of the newspapers or good periodicals or good academic journals?

CHAIRMAN BROWN: What you're adding is both the intelligence information, the synthesis of the two, and a distillation and some judgments. That really isn't what is usually meant. Well, sometimes it's what's meant by intelligence. But it's not what everybody thinks of as intelligence, which is a sensible reason for calling it a National Estimates Council.

MR. NYE: Right.


Mr. Vice Chairman?

VICE CHAIRMAN RUDMAN: I'm going to continue on that same line of questioning because I frankly think it's one of the more important areas that we've delved into, and we've heard from some very interesting witnesses. Let me put it just a slightly different way. I get the impression, and have for the last few years, even before I left this place, that we were getting more and more valuable information from open source. In many cases, we were getting it from other than conventional places, where you would expect to get that information. Of course, I, picking up on the Chairman's last comment, as far as I'm concerned, if the information is being brought to bear on any subject that's of interest to the Government, be it a glean through clandestined means or through open source, it's still intelligence. It's all put together, and somebody has to decide what it means.

I've got the impression from the bureaucracies which we have established, that there are a lot of places in this country, and probably in other countries, but let's stick to this country, that are extraordinary in various parts of the world. The information that they have given us in Intelligence Committee hearings, and other kinds of context, were frankly the most valuable we received. It should come as no surprise. There are people at the institution at which you now reside that have spent a lifetime with a whole series of graduate assistants and so forth working on one area of the world. The example that I made to my fellow Commissioners during our deliberative session yesterday was that you'll hear of something happening someplace in the world you never heard of, and that night you will not only learn more about what that place is and where it is, but the person you hear from on McNeil-Lehrer is from a university you've never heard of. It turns out he's the world's greatest expert, or she is the worlds greatest, on some narrow niche of information that nobody really knows much about. Now, you may disagree with that, but that is my view.

I have the feeling that we ought to do more and more; we ought to broaden what the Chairman and you were having a dialogue about. Maybe this office that would be part of the NIC, or the NIE. It would be a management agency to a large extent, contracting out with a lot of people around the country to focus on issues rather than bringing them in -- we're talking open source now -- let them do it where they are with the resources they have. And then it is up to the agencies of government to meld that information with information which obviously the academics don't have, or if they do have they probably shouldn't have it.

That really is just a long way of asking whether or not there is some way that we can do more of this outside of the Government, since so much of what we now call intelligence is truly open source compared to even 5 years ago and certainly 10 years ago.

MR. NYE: Well, I think the idea of just relying on the outsiders probably won't work. You need some place that pulls the outsiders together, and that distills their wisdom.

One of the things that struck me the first time I went into government in the 70's, having been an academic before I went in, was once I got into government I would find friends who would send me papers to help me, and I would get 30 or 40-page papers with lots of footnotes and so forth. They had no idea how useless that was. I mean, it might have the answer in it somewhere, but if my life was that I got to the office at 7:00 a.m. and had a half-an-hour to look through the part of the intelligence, the overnight intelligence, that my special assistant had underlined in yellow, and also get through some of the press clips and also return the most urgent phone calls before I had a 7:30 meeting with the Secretary of State, this 30-page paper was totally useless.

And the danger is that the academic is unconstrained by time. The biggest difference between government and academia, in academia the premium is to get the answer exactly right, A plus, even though it takes you 6 months or a year or longer to do it. In government, if you don't have the best answer possible on time it's worthless. Or putting it another way around, if you don't have a B plus by the time the Secretary meets with the President and you get him an A plus paper an hour later it's an F. And there's a total different world of the academic and the policymaker.

You need something like this National Estimates Council I'm talking about as a way to synthesize that; in other words, to get people who serve as a link between these two worlds. The National Intelligence Officer, along the lines that I describe, would be a person who has an understanding of how policy is working and what policymakers need and how to get the information crystallized and get it to the right person at the right time, but also who has a golden rolodex in his or her head of who out there in North Dakota or New Hampshire really knows the answer, who is the best person on this dimension of Islamic Fundamentalism in Tunisia, who may be in North Dakota or New Hampshire or wherever. And you need somebody who can pull that information together, crystalize it in a form that is useful to policymakers, and get it to policymakers at the right time. And you can't do that just by relying on bringing in outside academics.

VICE CHAIRMAN RUDMAN: Oh, I agree with that. Let me ask you, do you think we do that very well now, or do you think there's a lot of room for improvement?

MR. NYE: Well, we do it reasonably well. I mean, one of the things that I tried to do when I was Chairman of the NIC and which my successor Dick Cooper is doing who is a former colleague at Harvard also is trying to do now, is to bring in a mix of academics as National Intelligence Officers, so that in the position of the National Intelligence Officer will be people who have this web of contacts to the outside, and in addition to that to try to hold various conferences where you bring in the academics or tell your National Intelligence Officers to make sure they are constantly pulsing the outside network, but then to make sure that when they produce an intelligence estimate for the President -- I instituted what I called the President's Summary. It's supposed to be four pages. And instead of these long intelligence estimates of 20 or 30 pages -- we still publish those, because they are useful for the staffs, but you've got to keep it in a short, cryptic form for the top policymakers. And that takes sort of a linkage, somebody who has good views of the outside, but good views of the inside, and can put that into a particular crystallized form of knowledge.

VICE CHAIRMAN RUDMAN: Thank you very much.


DR. HERMANN: Dr. Nye, at the moment the U.S. has a dominant or a preeminent position in what I might call intelligence, reconnaissance, surveillance, and all that stuff. It is a distinguishing attribute of our sort of geopolitical position. My question to you is how important do you believe that is to our current and our future role in the world, as you see it?

MR. NYE: I think it's extremely important. Admiral Owens, Bill Owens, the Vice JCS, and I are publishing an article in the next issue, I think it is, of Foreign Affairs in which we argue that the information revolution has given the United States a particularly privileged position in power, that if we think in power terms we have to realize that the United States, because of being on the leading edge of the information revolution, is in a position of great international power.

Sometimes we think that you measure power by totaling up divisions or by just looking at GNP. I think that misses the extraordinarily important role that information plays. And among that information is this ability to get the type of surveillance, and not only surveillance but dissemination to the field that you described. So I would underline that as extremely important.

CHAIRMAN BROWN: General Allen.

GENERAL ALLEN: Dr. Nye, going on to your board of your estimative council, or whatever you called it, you've addressed it mainly in terms of its analytic kind of function. Could you imagine that when a requirement for a product arose and collection was started by whatever means, that there could also be a process which allowed a collection to be specified that might include contracts with academic or private think-tank organizations. Or is there some way of organizing in an effort to bring in more data for that particular product that would end up being a sort of collection function?

MR. NYE: Well, I think the National Intelligence Officer's role in collection is to report the policymaker's needs to the agencies that are responsible for collection, in the sense of clandestined collection, and the National Intelligence Officers now do that. They are, in this needs process -- they, because they have the most contact, or should have the most contact with the policymakers, are responsible to go back to the collectors of clandestined intelligence and tell them what's needed most, what's on top of the agenda of the policymakers.

But there is another form of collection which is open- source collection. And in that sense I think the National Intelligence Officers are extremely important, because when you call up Professor X at North Dakota State and say hey, I can't find out anything about Somalia, and yet you're an expert on Somalia, what can you tell me, that's the way in which I think they play a critical role in terms of open-source collection.

So I think they can play a role in both sorts of collection, but I would not want to see them get into the active business of clandestined intelligence collection.

GENERAL ALLEN: I understood, but my question was do you think we do that well enough. Or would this new council that would largely deal with open sources help that kind of collection from open sources?

MR. NYE: I think it could help. I think it could improve the process, because many of the people inside the intelligence community are aware of this need to reach outside, but the people who have spent their lives on the outside are often more alert to the distribution of information throughout the society and how it's distributed.

During the Cold War, for good reasons, because we had a maligned and talented enemy, the intelligence community developed a number of procedures essentially that cut it off from the outside. I mean, the various things, without getting into the merits or demerits of polygraphs or of pledges where you sign never to disclose anything you write for the rest of your life and the difficulty of getting in and out of the building, all these things were proper in terms of protecting against a Soviet enemy. But they also mean that there's a great difficulty in easy access to the outside. If you're dealing from Langley to a university or a think-tank. If you have this separate agency on F Street I think you'd have a much easier type of interchange.

In addition to that there is the fact that it's often easier for academics to have that type of interchange with a National Estimates Council than with the CIA or an intelligence agency. So I think for those reasons it could help more in the collection of open intelligence, open sources information.

A third reason I would say is that there is a danger of people in Langley, and I don't want to character this because there are some very bright naval people in Langley, thinking that information resides inside our four walls. In other words, if it isn't clandestinely, it can't be as important. And very often there is a danger that you would find particularly in economic intelligence, where I think there was an awful lot of trivia being collected, which anybody who had the sense to just sit down and read the financial times could have found out, and while we were risking any sort of possible diplomatic incident by collecting this stuff from other countries clandestinely when you could have gotten most of it by picking up the newspaper and reading intelligently, I think is explained only by the feeling that, well, if we got it clandestinely, somehow it's very important. And I think there is a psychological problem that we have to outgrow in that sense, and I think this type of body that I have described would help to also overcome that.

CHAIRMAN BROWN: Ms. Caracristi.

MS. CARACRISTI: I wonder how you would envision the relationship between the existing DI and your new organization. Would this have implications in at least the numbers of people that would be required in that world, or would some of those people be useful participants.

MR. NYE: I think there should be a division of labor. The DI should be the larger organization, should be the organization that's responsible for collecting current intelligence, assembling it, and distributing it. And I think the National Estimates Council, to give that new name to the NIC, should stay relatively small, flexible, with people on rotation. And I instituted the practice that an NIO should only stay for a maximum of 4 years. I think you want these people to turn over, bring in constant new blood from outside, but also from different agencies.

At the same time, I would like to see a large number of DI analysts rotating through the NIC, so that you keep good contacts between the DI and the NIC, or NEC, I guess we'll now have to call it.

CHAIRMAN BROWN: There already is a NEC.

MR. NYE: In this terminology. So the DI would be, in the division of labor, responsible for current intelligence, responsible for the basic analysis, would be the larger organization, and the NEC would be responsible for estimative intelligence, small, flexible, current turnover with a lot of overlap with staffs from other agencies from outside and from the DI, who would then work together to assemble a quite unique product on the estimative side.

CHAIRMAN BROWN: We have time for one more.

MR. FRIEDMAN: To follow up on the last question, Dean Nye, what, if any, functions of the NIC would you shuck in this new structure? In other words, would you change or slim down the NIE process? And second, would this National Estimative Council perform any quality control services for the DCI with respect to current intelligence. Would it point out things that perhaps were being politicized or where there might be, in their view, a need for competitive analysis? Maybe you could go into that a little bit.

MR. NYE: I think when I took over as head of the NIC we were producing something like 70 estimates a year. I regard that as too much, and I cut it down, and I think we were doing something like 40, 45. I think the estimates should be targeted on issues where there is a need on the part of the policymaker. In other words, not just what's new in Tajikistan, but unless somebody is about to go to Tajikistan or has a strong interest in it, we shouldn't be in that business. So we were on a track, at least when I was there, of trying to shuck off some of the extra work on estimates.

On the needs process, we instituted a process in which we did annual reports. Each NIO would to a report on his or her area which went to the DCI saying here's what we've heard from policymakers about their needs. Whether the newly formulated NEC should do that or not, I think should be left up to the DCI. In other words, the DCI may want to have that done elsewhere in the system. It strikes me as a useful function, if not formally in the current needs process, in some form. The NEC, which is going to have these constant interactions with policymakers, accumulates valuable information from the DCI's point of view, which is what's on the policymaker's mind. And finding ways to channel that back to the DI and to the DCI is extremely important.

Let me make absolutely clear, it would be a big mistake if you set up on F Street a little CIA or a little DI. That's just not my model at all. I'm talking about two overlapping organizations with a division of labor which complement each other.

MR. FRIEDMAN: Really, a reengineered NIC which you've moved away from Langley and somewhat changed its functions, if I hear you right.

MR. NYE: Right. When I took over the job, or was offered the job in early '93, there was a plan to move the NIC to F Street, which in the budget down-sizing fell victim. I can't now remember the cost. It might have been something in the range of $20 million, which is money. But in other words, to try to bring the building at F Street up to the standards that were necessary and so forth, it was going to be an expensive move. And I think in that budgetary climate the feeling was it was too expensive. I think, in retrospect, that that's too bad that that decision went that way.

CHAIRMAN BROWN: Mr. Dewhurst, we do have time for one more question.

MR. DEWHURST: Mr. Secretary, good to see you again. One of the half-a-dozen areas that has concerned a number of us, including myself over this past several years, is proliferation, particularly to the extent that it may touch terrorism and affect the security of the United States. This is a subject that you have been interested in for a long time. Could you share with us your views on how good of a job the Community is doing in the field of monitoring proliferation, whether it's from overt state, clandestined sources, and what we might need to do better in the future?

MR. NYE: I looked at this proliferation issue from both sides of the table, both inside intelligence and as a consumer when I was in Defense, and earlier than that, in State. I think this is a high priority, and it's something we've done quite well. I mean, I was often pleased by the quality of the information that we received, and this is an area where some of the clandestined information was just extremely important. This is an area where open source covered some things, but in which the clandestined is extremely important, and which the proportions I described would be heavily weighted on the clandestined side.

The nonproliferation center that Gordon Naylor was running I think did good work. John Deutch had a strong interest in this, and when he was still Deputy Secretary of Defense he ran a committee on which I sat which looked at the priorities for collection and for action by both Defense and the CIA.

I think we can always do better, but I am impressed by how well we have done. I mean, there is some information that we have on proliferation which is truly important. And if somebody says that we don't need clandestined collection of information in a post- Cold War period, there's no overriding threat, no Soviet Union any more, I think there's a recent article in Foreign Affairs that says something along these lines, that's not my experience at all with the information that we've got on proliferation. We got some truly first-rate information from a whole variety of sources, but including some from clandestined humint sources.

CHAIRMAN BROWN: Thank you very much, Dean Nye. I think you have contributed a good deal to our understanding of these issues, and you make, I think, a very interesting suggestion about a National Estimates Council. Thank you, again, for appearing here.

MR. NYE: Thank you, and I wish this commission well, because it's got a terrifically important job.