As a young staffer for the first Senate investigation of intelligence in the 1970s—often called the Church Committee after its chairman, Senator Frank Church (D–ID)—I was briefed by Clair George on CIA spying in Asia. George, then chief of the Agency’s East Asia Division, later headed the CIA’s spymasters in the Directorate of Operations (DO), when he was convicted of lying to Congress in the Iran-Contra affair. At one point in my meeting with him he argued forcefully: “I’m in the business of producing spies; if I wanted to produce widgets, I’d be in private industry.”

I was impressed by that statement at the time. Most of my studies had been on the diplomatic side of U.S. foreign engagement, where virtually no one thought he produced much of anything. I later learned, however, that George was wrong: His real business was not producing spies, or even producing information. His real business was producing useful insights in the heads of national officials charged with making decisions.

That understanding is at the core of Richard Betts’ slender and insightful book, Enemies of Intelligence. Betts has given us essentially a long essay about the connections between intelligence and policy, with excursions into collection, the tradecraft of analysis, and the collision of privacy and security in domestic intelligence. To do so, he draws on a rich trove of historical materials. Many of his examples are from the Vietnam era and so may seem less immediate to younger readers, but the range of this work and his comparisons with other intelligence services, especially Israel’s, enrich the book throughout.

Richard Russell’s Sharpening Strategic Intelligence is also readably slim yet covers a lot of ground—and covers it carefully. It is a complement to Betts, for Russell’s focus is analysis. His ostensible target is the CIA, the agency where he worked for nearly two decades as an analyst, but his real subject is intelligence more broadly construed. By the same token, while his title addresses “strategic intelligence”, the litany of failures he reviews also encompasses intelligence that usually would be regarded as tactical or operational, such as wartime targeting or assessments of bomb damage produced by U.S. air attacks. In part, this is because Russell defines strategic intelligence very broadly as “the use of information, whether clandestinely or publicly acquired, that is synthesized into analysis and read by senior-most policymakers charged with setting the objectives of grand strategy and ensuring that military force is exercised for purposes of achieving national interests.”¹

Neither author ventures far into the post-9/11 challenge of collecting more domestic intelligence on citizens and resident aliens. Betts’ penultimate chapter does, however, make a very useful argument about intelligence collection and civil liberties. For him, the public debate mostly gets it wrong. There is no collision or need for any trade-offs between liberty and security. The real balance at issue concerns security and privacy, “the one aspect of liberty that inhibits the government’s acquisition of knowledge.” There is no need to compromise freedom of speech, religion, political organization and, especially, due process. By Betts’ lights, privacy, or “personal secrecy”, is important, but not as vital as freedom from being jailed without trial. “The focus of concern should shift from limiting government acquisition of private information to measures for strictly limiting the use of that information”, he writes. That will require oversight mechanisms with teeth.

Intelligence and Policy

“Politicization” has become a fighting word in the world of intelligence, connoting analysis bent under pressure to suit the preferences of senior bureaucratic bosses or politicians. Betts has had to rethink his own approach to politicization, especially in the wake of the October 2002 National Intelligence Estimate (NIE) asserting confidently that Iraq had weapons of mass destruction. My experience, like his, had led me to be skeptical of “politicization” epithets. Intelligence analysts live and die by their written analyses. Hurt feelings and damaged egos being what they are, a reviewer’s criticism could easily seem like politicization to an aggrieved party. Betts quotes Mark Lowenthal, a man with considerable experience both outside and inside intelligence, as saying that sometimes those who cry “politicization” are simply the “‘losers in bureaucratic battles.”

In fact, the bigger concern is, to Betts’ mind (and mine), the irrelevance of intelligence. “The best intelligence is useless if those with authority to act on it do not use it”, writes Betts. Or as he put it in his Foreign Affairs post mortem on the post-September 11 reforms of the intelligence community: “The typical problem at the highest levels of government is less often the misuse of intelligence than the non-use.” Analysts paid a high price for appearing to “get on the team”, “hew to the line” or otherwise commit the sin of becoming “politicized.” However, they paid no comparable price for being irrelevant. Too often, intelligence products seemed to me to answer questions no one was asking. The questions were either ones that interested the analysts or were “safe” questions, amounting to the equivalent of “whither Germany?” Policy virtually never asks “whither Germany?” Perhaps it should, especially for the purpose of long-term planning, but that planning is as rare as hen’s teeth.

For Betts, the challenge is precisely to politicize intelligence in the non-pejorative meaning of the term. That means bringing intelligence into the realm of politics, to questions actually asked, and then trying to sustain the distinction between the bad, corrupting analysis, and the good, honest analysis in ways that “engage policymakers’ concerns.” He quotes Robert Jervis’ pithy remark: “It is . . . easy to keep intelligence pure when it is irrelevant.”

So how does one achieve positive politicization? The proposal to fix the term of the director of national intelligence in the manner of the FBI director is not a good idea, says Betts; the reason is that rapport with the president is of paramount importance for a DNI, and no president should have to tolerate a predecessor’s selection. He does, however, suggest several other ways of providing insulation from undue political pressure—for example, by appointing a DNI from the opposition party or a retired elder statesman, or by creating oversight institutions comparable to the Federal Reserve Board and the Government Accountability Office.

Betts rightly identifies the October 2002 National Intelligence Estimate on Iraq as a failure more of collection than analysis. To be sure, the analysis was dead wrong about Iraqi WMD programs, but “it was the proper estimate to make from the evidence then available.” The real failing of the estimate was that its managers succeeded where others of us who have managed the estimate process have failed: They kept weasel words out the estimate, so that it sounded much too certain of its conclusions given the paucity of evidence. In that sense, former Defense Secretary Donald Rumsfeld was wrong when he said at a 2002 press briefing: “The absence of evidence [that Iraq has WMD] is not necessarily the evidence of absence.” In a Bayesian sense, each day the UN inspectors didn’t find WMD ought to have weakened confidence in the judgment that Iraq must have them. But the workings of cognitive dissonance know nothing of Thomas Bayes and his theorem.

Both Betts and Russell know that mindsets are the root of most mistaken intelligence analyses. They both know, too, that the hardest analyses to challenge are not the cases of “mirror imaging” opponents’ intentions (what would we do in their shoes?). Rather, the hardest cases arise when analysts do a good job of assessing what a target should do in its own shoes, and then it goes off and does something different. The classic case is the Soviet deployment of nuclear missiles in Cuba in 1962, which, even with all the benefit of hindsight, looks like an insane strategic gamble by normally conservative Soviet leaders.

Betts calls that “exceptional thinking” versus “normal theory.” Both he and Russell quote Douglas McEachin, the longtime CIA Soviet analyst and later head of the Directorate of Intelligence (DI), the analytic arm of the agency. Russell quotes McEachin calling those exceptions the “dumb move. . . . When some analysts did try to make the case for the dumb move, they were also characterized as dumb.” In Betts, McEachin repeats the old CIA joke after the December 1979 Soviet invasion of Afghanistan: “The analysts got it right. It was the Soviets who got it wrong.”

Jervis, again quoted by Betts, puts it most sharply: “It is particularly difficult for analysts to get it right when the truth is implausible.” Judge Richard Posner put implausibility in somewhat different terms in his review of the 9/11 Commission Report, less as an obstacle to imagining than to acting. For him the ultimate cause of the 9/11 failure was banal: “It is almost impossible to take effective action to prevent something that hasn’t occurred previously.” Because exceptional thinking is “dumb”, ostensible experts are all the less likely to think of it. What is required is a “seasoned generalist” with enough stature in the process not to be dismissed, whether as dumb or merely eccentric.

Diagnosis and Prescription

Neither book provides much to quibble over, but honest observers may take other views on what are genuinely difficult issues. Russell’s overall diagnosis is damning, and the recommendations he makes are solid ones, but there is perhaps something of a gap between the two. Both sides of the CIA—the DO, now the National Clandestine Service (NCS), and the DI—are bureaucratic and stuck in old practices, ones that, by Russell’s account, didn’t work very well to begin with. His portrayal of the DI as an almost anti-intellectual managerial culture accords with my own impressions from inside. Many DI officers have first-rate minds, but the organization feels more like a newsroom than a university. It rewards getting hot tidbits to policymakers quickly more than it does deeper analytic effort. Russell’s advice is to hire more Ph.D.s and find ways to nurture deep expertise, while emulating private corporations by eliminating layers of middle management.

Russell also endorses Robert Baer’s suggestion of creating different tiers of security clearance in order to bring into service a wider range of both Americans and foreigners. His other guidance is hortatory: Simply hiring more people won’t help, and saying that creating spies takes time is a cop-out. The CIA can do better, and it should focus on “the most critical issues facing U.S. national security decision makers.” Well, yes…

Russell’s most specific recommendations, also sensible ones, concern not the CIA but rather the National Intelligence Council (NIC), my old home. Russell would have it shun its newly acquired role overseeing the President’s Daily Brief (PDB), the CIA’s crown jewels of current intelligence. Instead, the NIC would be beefed up (it is already being expanded) and would reach out still more to the academy and think-tank world. It would become the focal point for strategic intelligence. In parallel, Russell seconds the 2005 Robb-Silberman WMD Commission recommendation of an intelligence-sponsored non-profit research institute “to serve as a critical window into outside expertise for the intelligence community.”

advertisement

Betts, also the author of the classic Surprise Attack, has studied intelligence and been in government long enough to be skeptical of post mortems and exhortations to achieve dramatic improvement. Post mortems tend to be rare and futile, more often focusing on finding villains to blame than on improving practice. Even when they get beyond “who shot—or who missed—John?” questions, they still are limited in what they can achieve because they focus on only a handful of incidents that cannot reliably be generalized to different circumstances. They tend to assume that if analysts did x and failed, then doing not-x (or anti-x) would have produced success, and would do so in future circumstances akin to those examined. But this is not true. They also tend to be either superficial, if they are stocked with wise people who are nonetheless amateurs in the esoterica of intelligence, or biased, if they are stocked with experts even remotely associated with the failures under examination.

Recent major investigations by the 9/11 and WMD panels were both impressive by the standards of such ventures, but even they could not entirely escape those limits. Betts cites instances in which both reports advanced a recommendation, then countered it in effect, all within a few pages. The 9/11 panel did an unprecedented job of remaining in Washington after its report to hector both Congress and the Executive Branch into action. Yet as Posner and others have pointed out, its recommendations did not flow directly from its diagnosis. And even its wonderful, awful narrative of what went wrong has something of the historian’s fallacy about it: Because we know how the story ended, the clues along the way seem obvious, much as flight schools became instantly interesting after the World Trade Center went down.

Betts is a fatalist, the Eeyore of intelligence analysts. In his view, intelligence failures are inevitable, and reform measures will have side effects and unintended consequences that “sometimes produce benefits, but . . . always have costs.” Dour as his assessment may be, it isn’t easy to show that he’s wrong. Many recent favorites, like trying to centralize more power at the center—now the DNI—have been looked at and shunned in earlier cycles. In Betts’ words: “The reasons for this failure to follow through with recommendations for radical change have something to tell us about the prospects for progress from the 2004 legislation” that created the DNI. Again and again, when he applauds initiatives, it is wanly as “more good than bad.”

“What can be improved easily helps marginally”, says Betts, “while what can help more than marginally cannot be improved easily.” In contrast to Russell, Betts applies that second clause especially to U.S. espionage against targets like al-Qaeda. The dilemma at the center of the recent proposals is that major failures like Pearl Harbor or 9/11 seem to argue for more centralization, which is the direction of the 2004 legislation. On the other hand, day-to-day concerns about the way government works tend to favor the other tack, enlarging pluralism and redundancy, and cultivating the healthy competition they engender.

Betts addresses Russell’s call for cutting layers of bureaucracy, and even there he is cautious: “If lots of chaff is tossed out, some wheat is bound to go with it.” Indeed, the failures of both 9/11 and the Iraq NIE—missing dots that might have been connected and jumping to strong conclusions on weak evidence—can be read as requiring more people doing more things, including adding more checks and balances, not fewer.

Betts ends where he begins, with the connections between intelligence and policy, observing that there ought to be “some logical relation between pessimism about how well knowledge can be summoned by national security policy and the aggressive ambition with which that policy should try to control the course of events everywhere in the world.” In that injunction, he echoes our common mentor, Richard E. Neustadt, who wrote almost forty years ago in Alliance Politics that the United States ought to limit its demands on foreign governments, even allied ones, to “outcomes which do not depend for their achievement on precise conjunctions of particular procedures, men and issues.”

In teaching intelligence professionals, I turned this guidance into cookbook language: The less you understand about them over there, the better you should understand you over here, the capabilities and proclivities of your institutions and your political masters. That guidance seems to me all the more apt in this season of reshaping intelligence.

1.
In the spirit of full disclosure, I should say that Richard Betts and I have been colleagues and friends since graduate school, and we served together on the Church Committee. I know Richard Russell only through his writing, but his book is generous to me and my home institution, the RAND Corporation.