PrintPrint EmailEmail ShareShare CiteCite


Betts: Don't Expect Perfection from Intelligence Agencies

Interviewer: Bernard Gwertzman, Consulting Editor
Interviewee: Richard K. Betts, Adjunct Senior Fellow for National Security Studies
June 21, 2004


Richard K. Betts, director of War and Peace Studies at Columbia University and a widely published author on intelligence matters, says the failure to prevent 9/11 and misjudgments about Iraq’s weapons of mass destruction were “not among the worst or most unusual” intelligence mishaps. “The problem with most intelligence failures is that they are obvious after the fact. But before the fact, there are lots of reasons that a mistake in judgment gets made,” says Betts, an adjunct fellow at the Council on Foreign Relations.

Betts views proposals to create a new director of national intelligence skeptically. “A new director of national intelligence could have a revolutionary impact if he or she were given authority to direct the missions, the priorities, and the activities of all of the 15 intelligence agencies throughout all the departments of the government,” he says. “But I will believe that when I see it.”

He was interviewed by Bernard Gwertzman, consulting editor for, on June 21, 2004.

Other Interviews

How do recent intelligence failures— the 9/11 attacks and erroneous assertions about Iraq’s weapons of mass destruction (WMD)--compare with intelligence missteps of past years?

As bad as they are, they’re not among the worst or most unusual. The problem with most intelligence failures is that they are obvious after the fact. But before the fact, there are lots of reasons that a mistake in judgment gets made. The September 11 case is one where if we had connected all the dots in a way that now seems logical, it could have been prevented.

There was enough information available in different places that, if it had been shared in a more efficient way and if people had been smarter in making deductions from it, conceivably this could have led to some sort of action that would have interfered with the plot. But those are two very big “ifs,” and it’s entirely possible that even with more efficient sharing of information, the critical leap to an investigation of flight schools might not necessarily have been made. But at least there was a chance. This now looks in hindsight like something for which there’s no excuse for missing, but I think it was a lot harder, obviously, to see that before the fact.

On the weapons of mass destruction, it appears that the biggest scandal was not that intelligence was misrepresented or misunderstood, but that there was so little positive intelligence. Circumstantial evidence that the Iraqis had weapons of mass destruction was overwhelming, but what it looks like now is that there wasn’t much beyond the circumstantial evidence. And the circumstantial evidence was how they had obstructed inspections during the 1990s, played cat and mouse with UNSCOM [United Nations Special Commission], and blocked the inspectors so much that they had to withdraw. That didn’t make much apparent sense unless they were trying to hide the weapons we knew they once had. There’s no dispute that the Iraqis did, before the 1991 war, have chemical and biological weapons of some sorts and of some quantity. What it now appears happened is that Saddam Hussein may have been telling the truth when he said they were destroyed after the war.

Could you give some examples of worse episodes of intelligence failure?

The sad but common fact is that surprise attacks often succeed despite warnings of enemy preparations. They succeed because the warnings are ambiguous, or misinterpreted, or get lost in the noise and confusion of diplomatic maneuvering, or are simply disbelieved.

Recurrent mistakes include dismissing warnings because an attack seemed to be an irrational choice for the enemy; the “cry wolf” problem, when several false alarms make policymakers take subsequent warnings less seriously; guessing correctly that an attack will occur, but guessing wrongly about exactly where, when, or how; having warnings held up too long in the communication chain between operators in the field, processors in the intelligence organization, and policymakers; judging that an attack is likely, but waiting too long for more information before deciding what to do; or being misled by enemy deception operations.

For reasons like these, major countries have fallen victim to surprise numerous times in past decades, for example: the French, British, and Russians before the German blitzkriegs of the 1940s; the Arabs in 1956 and 1967; the Israelis in 1973; and the United States before Pearl Harbor, the invasion of South Korea, the Tet Offensive in Vietnam, and the Iraqi invasion of Kuwait in 1990. If anything, U.S. and allied intelligence has probably done better overall in breaking up terrorist plots in the past 10 years than governments have done in avoiding conventional surprise attacks in the past century.

It looks, to some people at least, that Bush administration officials pushed to have intelligence that would support the case for war.

I’m sure they wanted the intelligence community to come up with as good a case as it could. But I think most people in the intelligence community really believed that Iraq had weapons of mass destruction because of that circumstantial evidence. There’s a powerful tendency for intelligence estimators to assume that an enemy is going to behave rationally— at least, what seems rational in our terms. And the Iraqi behavior didn’t seem rational in our terms. We would have thought if they really didn’t have any, they would have let the inspectors go anywhere they wanted forever in the 1990s, rather than stonewalling and appearing to hide things.

Former top U.S. weapons inspector David Kay has theorized that Saddam Hussein wanted people to believe he had these weapons in order to keep his country together.

Now we can figure out the rationales the Iraqis or Saddam Hussein personally may have had for pretending, implicitly, to have the weapons even while they were officially denying it. Once you see what appear to be the facts, then you can find some rationale that makes sense: that he was trying to play it both ways to get the deterrent benefit of having us think he had the weapons while getting the diplomatic benefit of denying it. Before we had the current evidence, or the lack of evidence, of weapons being in Iraq, that would not have seemed very logical, so that’s why I think intelligence estimators honestly agreed with the administration that it was almost certain [Iraq had WMD].

What is surprising is that there seems to have been so little even fragmentary information to support that deduction. I assumed before the fact that we didn’t know everything about where these weapons were, but we had enough clues to find some parts of their program pretty quickly or tell the U.N. inspectors, when they were finally let in right before the war, where to look. And the fact that that wasn’t true, I think, is one of the bigger surprises of the intelligence failure.

You warned in 2003 that Iraq might use WMD.

Yes, I believed, like everyone else. I believed before the war that they had them, again, on the basis of that circumstantial evidence, although I also, perhaps naively, assumed that there was more positive evidence on the inside than I had access to.

Is U.S. intelligence-gathering as flawed as some of its critics say?

It may sound like a cop-out, but I think the question is, “Is the glass half full or half empty?” [Intelligence-gathering] is in terrible shape if your standard for what is good intelligence is never missing an important event or never failing to predict an enemy attack before it occurs. It’s in good shape if you think the odds are stacked against intelligence to begin with and if [intelligence officials] do it right three out of four times or even two out of three times, that’s doing pretty well. What should our standard be? Realistically, it should be more the latter; that is, when you’re dealing with enemies who are trying to outwit you and who know a lot about what you do to get information on them and who try to adjust their behavior to try to prevent you from doing that, then finding out about a majority of plots, even if you don’t find out about a significant number of others, is doing pretty well.

The intelligence community gets very little credit on the outside for all of its successes. If you go back through the press and you look on the bottom of inside pages of The New York Times or The Washington Post every few months, you will see a plot being reported broken up as a result of getting information on the plotters before they could actually execute it. But that doesn’t register with most people. It’s sort of taken for granted: “Great, that’s what they’re supposed to do.”

But I liken it to the problem of improving a batting average. If somebody is trying to get the ball past you, even if you are a tremendous hitter, they are going to strike you out occasionally. That’s not something that people like to hear. It sounds defeatist; it sounds like it’s not asking enough of our public servants. I think it’s realistic. Things could always be done better. Any batter who strikes out realizes he might have adjusted his stance, he might have swung at one pitch rather than taken it, he might have done this or that [in order to] avoid the problem. I think to believe an adequate and properly functioning system will never have a serious error is just not giving enough credit to the enemies who are working pretty hard to outwit you.

You have written about the importance of human intelligence. In Iraq, it doesn’t seem as if the U.S. forces have much solid information on where the violence is coming from or how to prevent it. Is that because they are foreign occupiers?

For one thing, we can hope that some of the intelligence is better than we know on the outside. But, assuming that it’s not and we don’t have very good sources on various parts of the resistance, I think it’s due to a number of problems. First, we are an occupying power, which limits the number of people in the country who are going to jump at the chance to inform on thugs who are plotting against us.

Second, even the ones who want to help us, if they have information, have reason to fear for their lives, and it’s pretty obvious that they can’t count on us to protect them. If we had a better a security situation in Iraq and the threats to people who collaborate with us were under control, we probably would get a lot more cooperation and a lot more sources pouring in.

Also, we don’t know who to trust when we do get human intelligence. The constant, age-old problem with human intelligence, which really is something that all the James Bond mythology and spy movies we grew up on do not convey, is that human intelligence sources, when they do exist, are widely distrusted because they very often prove to be either flaky or dishonest or working both sides of the street. It’s hard to tell if a source is real and giving you the truth, rather than trying to play you for a sucker. In Iraq, we probably don’t have many sources to begin with; our confidence in the ones we do have is limited.

And finally, things are fast moving. If you don’t have many people who are your own trusted agents, who speak the local language, and who have contacts built up over the years, what you are going to get comes in bits and pieces and may come in too late to give you the kind of intelligence you need to avoid things like bomb plots.

Some have called, as you mentioned in a recent Foreign Affairs article, for the appointment of a director of national intelligence. I don’t see how that position would differ from the director of central intelligence (DCI).

Bingo. You broke the code. If people really wanted to give authority over all these intelligence agencies, they do not need a new grand design to do it; they could have given that authority to the DCI years ago. And the reason that he doesn’t have as much authority as some critics now believe he should have is that it has always been resisted for reasons that are both good and bad.

Good reasons: there are benefits to having independent centers of collection analysis in the community; also, each department that has its own intelligence unit has its own specific needs and needs to have a unit that is very responsive to them. Bad reasons: all the normal bureaucratic turf wars and jealousies that go with any complex organization like the U.S. government. A new director of national intelligence could have a revolutionary impact if he or she were given authority to direct the missions, the priorities, and the activities of all the 15 intelligence agencies throughout all the departments of the government. But I will believe that when I see it. The risk is getting legislation that creates a new director of national intelligence with a fancy title and a vague mandate to exert more authority but who doesn’t have the actual last word on what all these agencies will do. It’s very hard for me to believe that the departments— the State Department, the Treasury Department, the Energy Department, and most especially, the Defense Department— will allow that sort of complete authority over a unit within their department to be given to somebody outside of their department.

How does the current system work in practice? The DCI, of course, runs the Central Intelligence Agency. He also chairs interagency meetings on intelligence, but does he have any authority over the Pentagon’s huge operation?

He has a lot of control over budget allocations and organizing programs because of previous legislation and because [the DCI’s] authority has increased over the years. He doesn’t have complete control over everything the National Security Agency or the Defense Intelligence Agency or the Bureau of Intelligence and Research in the State Department— everything that they do from day to day. He does have a lot of influence over resources, but it’s not complete control. If you want to give him complete control, you might as well have a Department of National Intelligence, just like the departments of State and Justice, in which all the intelligence agencies are taken out of those other departments and put into this new department with one secretary of intelligence. But [even though] many are promoting this big reform of a director of national intelligence, I don’t know anyone who’s actually advocating taking these units out of the other departments.

Who have been the best directors of central intelligence? And who would make a good one now?

There’s no one answer to the last question. Because the job is so complex and involves so many different kinds of skills, no one individual is going to have all of them, and therefore, it depends on what you think is the highest priority. One could make a case for a number of things: a business manager, a consummate analyst, a spy master. Choices among [those] priorities would probably lead you to different choices of directors.

One of my favorites is James Schlesinger, who was only in the job for a few months [in 1973], back in the Nixon administration. I like him for parochial reasons. Being an academic, I put high priority on the importance of analysis, and he was about the only professional analyst we had in the job. He also had a lot of administrative experience in several areas of the national security bureaucracy— the Atomic Energy Commission, the Bureau of the Budget, and elsewhere— so he brought a breadth of experience and sensitivity to intelligence problems from the government as a whole and from within the intelligence community. That, I think, is an ideal combination of analytic skills and managerial and policy background.