Thoughts on the SSCI Report, Part III: The Program’s Effectiveness

By Benjamin Wittes
Sunday, December 28, 2014, 11:32 PM

Having started this series of posts by focusing on the aspect of the SSCI's report on which the committee majority is strongest---the program's brutality---I want to turn now to the aspects of the committee's work on which, to my mind at least, its report is less persuasive.

An enormous amount of the committee's energy is devoted to demonstrating that the interrogation techniques used in the program were not effective, and the committee certainly highlights areas in which the CIA stated too confidently that coercive interrogation played a key role in achieving positive outcomes to which it may or may not have contributed. But the committee attempts something more ambitious than simply arguing that the CIA overhyped the value of detainee interrogations---and the value, in particular, of coercion in those detainee interrogations. The committee attempts to argue that every single one of the intelligence successes the CIA has attributed to the program are in error---that is, that coercive interrogation led to none of the important wins (foiled plots or captures of important Al Qaeda figures) which the CIA has claimed as its fruits.

On this point, at least in my judgment, the committee's report is not persuasive.

Let's pause a moment here and consider why this is an important claim in the first place. After all, the committee would have been on safe ground had it merely pointed out cases---and there are many---in which the agency made representations it could not possibly know to be true about the effectiveness of its interrogation techniques. The agency on numerous occasions claimed that it had obtained life-saving information by means of these interrogation tactics, and that this information was not obtainable by other means. Even it now concedes that it is unknowable what sort of information it would have gotten by means less coercive than the ones it used---in other words, that we cannot assess the marginal value of coercion to the agency's intelligence gathering efforts. But the committee chose to go further and, like the agency itself, try to assess coercion's marginal value. Unlike the agency, it finds the value to be essentially zero.

The reason this is an important finding---if it's right---is that it eliminates all moral complexity from the discussion of the RDI program. No longer are we balancing the horror of torture---or something very close to it, if you prefer---against great intelligence gains, or even small intelligence gains, or even, as the CIA would now have it, major intelligence gains in which we cannot know the precise role coercion may have played. We are instead balancing the horror of the violence and brutality against exactly no gain. That makes the equation a very easy one indeed. It makes the past a painful one, in which we did great wrong and achieved nothing in exchange, but it also makes the future a bright and clear one: You get better results, and without the moral stain, if you just say no.

In his preemptive designation of me as a "torture agnostic" the other day, Steve Vladeck argues that:

as has been well-covered elsewhere, agnosticism as to torture’s efficacy is utterly irrelevant to torture’s legality. The law doesn’t brook an “efficacy” exception for torture any more than it brooks such a carve-out for other crimes. We may not care, in such cases, about the legal consequences of our conduct, but that’s not because the fact that we deem something to be pragmatic somehow transmogrifies the illegal into the legal.

Steve's point is correct, but it is misdirected. It was the committee majority, not any of the program's defenders---much less we poor agnostics---that staked out bold new ground on the question of the RDI program's efficacy. It was the committee majority that was not content to argue merely that the program was immoral and illegal but insisted on arguing that it was entirely ineffective as well. That bold statement requires an evaluation.

In any event, there's another reason the question of effectiveness remains important: Not all coercion is torture or cruel, inhuman, and degrading treatment. Leave aside for the present the endless debate over where the legal line in fact lies. It is certainly the case that the Army Field Manual, by design, does not come close to approaching it. Which is to say that some degree of coercion can be legal, and it therefore remains an important question whether it does any good---or whether it does harm---for intelligence gathering.

The committee majority purports to be assessing whether the "CIA's representations that its enhanced interrogation techniques produced unique, otherwise unavailable intelligence that led to the capture of specific terrorists and the 'thwarting' of specific plots were accurate." In fact, the committee's undertaking is more ambitious that that. It attempts to debunk every one of 20 incidents in which the agency claimed significant operational gains as a result of coercive interrogation within the RDI program. And it draws conclusions far more sweeping than the mere claim that the program did not produce unique, otherwise unavailable intelligence that led to important captures or the thwarting of specific plots. The committee's very first finding, for example, states that "The CIA's use of its enhanced interrogation techniques was not an effective means of acquiring intelligence or gaining cooperation from detainees."

Indeed, a fair reading of the committee majority report would have to assess the net national security impact of the program was highly negative. Because not only does the committee not accept a single one of the agency's claimed successes from its interrogations, it notes that the agency got a lot of bad information from detainees, and the majority also argues that the program "complicated, and in some cases impeded, the national security missions of other Executive Branch agencies." What's more, it also "damaged the United States' standing in the world, and resulted in other significant monetary and non-monetary costs."

All of which puts a heavy premium on the question of whether, in fact, the program's intelligence harvest was really as barren as the committee claims---or whether, perhaps, we "torture agnostics" might be wise not to assign a value to a single variable in a highly complicated, multivariate historical equation.

The committee is persuasive, to my mind, on one point related to the program's efficacy---beyond the CIA's having overhyped it: There is no smoking gun case in which a detainee went from obstinate silence to yielding up the unambiguous crown jewels after a period of coercion. There is no case as simple as the stories the CIA and the Bush administration floated about detainees going from ominously promising future attacks to revealing how to stop them and where to find each other once they had been "broken." The reality, the committee shows, was far messier than that. And the agency should be held accountable for creating too straight a line between coercion and the positive outcomes it achieved.

That said, the CIA had a lot of very big successes, and the strategic interrogation of detainees was inextricably interlaced with the history of---and interrelation between---a great many of them. The committee's insistence that all of the benefits are entirely severable from all of the coercion is innately implausible and requires a number of fallacious analytical leaps.

The first of these is downplaying the importance of corroboration. If a detainee said something and CIA already had that piece of information somewhere in its files, the committee regards that piece of information as merely corroborative and dismisses its importance on that account. Leave aside that the earlier piece of information may not have been noticed, or that its importance may not have been appreciated until a high-value detainee mentioned it too. So, for example, the committee dismisses the idea that the Abu Zubaydah interrogation led to the capture of Jose Padilla, among other reasons, because there was already reporting on Padilla in the CIA's possession. But the CIA responds, quite reasonably, that it had no reason to take this reporting especially seriously until Abu Zubaydah described Padilla too. It gets, after all, a lot of suspicious traveler reports. It only gets a very small number that match a person a high value Al Qaeda operative says is coming to launch a major attack.

The second fallacy is the confusion of what could have been with what, in fact, was. In many of these cases, the agency appears to have had in its possession information that with perfect processing and perfect analysis would have obviated the need for reliance on the detainee reporting. In fact, however, that is only clear in retrospect. And in real life, the detainee reporting---and the analysis of it---was, in fact, central to how the agency put two and two together. So, for example, the committee notes that Abu Zubaydah's identification of the elusive "Mukhtar" as Khalid Sheikh Mohammed merely corroborated a report already in CIA's possession. Now it's certainly possible, in a different, better world, that analysts would have noticed this report and that it would have settled the discussion. But, the agency responds, as it turned out, analysts "simply missed" the earlier report. Whether or not that earlier report would have settled the matter had it been noticed is, in the agency's judgment, an open question. "It is clear," as it happened, however, "that CIA only made a definitive determination that KSM was 'Mukhtar' after receiving the information from Abu Zubaydah."

Perhaps most importantly, the committee's dogged commitment to debunking all cases in which the agency claimed benefits from coercive interrogations lead it to take some unfortunate liberties with important facts.

The most significant of these involves the Abu Zubaydah interrogation, which nobody doubts yielded a good deal of ripe fruit. The committee argues repeatedly, as have critics of the CIA's program for a number of years now, that all of that fruit came from an early period of the interrogation, during which FBI agents used rapport-building techniques. The later period, which followed the initiation of the RDI program and involved a host of rough techniques (including waterboarding), in the committee's telling, yielded very little of value. So, for example, Abu Zubaydah's description of Padilla and his identification of KSM as Mukhtar both took place while the FBI was playing nice with him. At various points, the committee uses the contrast between the productivity of the earlier rapport-building phase of the interrogation and the non-productivity of the brutal phase as evidence that the enhanced interrogation was not effective, even that it eroded intelligence-gathering effectiveness.

The trouble, as both the CIA and the minority point out, is that the nice phase of the interrogation wasn't all that nice and included sleep deprivation techniques that were later incorporated into the RDI program. Indeed, the rapport-building phase of Abu Zubaydah's interrogation included a kind of tag-teaming of CIA and FBI interrogators, in which one side played nice and the other decidedly did not---making it impossible to draw a sharp line between a non-coercive phase and a coercive phase. The minority report claims that during the supposed rapport-building period, Abu Zubaydah was subjected to extended sleep deprivation, sensory deprivation, a liquid diet, and nudity. The CIA describes it a little less dramatically, as merely involving sleep deprivation. But either way, the point is that we shouldn't hold the take from this period out as the fruits of rapport-building interrogation alone.

My point here is not to insist that the program "worked" or to assert---in contrast to the committee---that there's a clear relationship between rough techniques and outcome. I do not know what would have been doable had the CIA used rapport-building alone, much less had the agency not been in the detention and interrogation business at all. My point, rather, is that the committee is arguing more than the facts will bear when it claims to know that the gain was zero.

In the end, I agree with the CIA that the answer to the efficacy question is unknowable---at least without deploying study methodologies we are unlikely ever to use. In some warped alternative universe, one could imagine, for example, a randomized trial in which a given detainee would be randomly assigned to rapport-building or enhanced interrogation, and the agency could determine, over time, which approach produced more useful intelligence. In our world, however, I cannot imagine anyone actually conducting such an experiment, which would be as vulnerable to ethical allegations of randomizing people onto torture as it would to---all too literally---rolling the dice with national security interrogations. But in the absence of such a study, we need to learn to live with a degree of uncertainty on this point.

I don't believe we're going to make a final definitive efficacy judgment based on cable traffic.