Back in September, I wrote the following:
Imagine you were a high-level decision-maker in a clandestine intelligence agency. Imagine that you had played by the rules Congress had laid out for you, worked with oversight mechanisms to fix errors when they happened, and erected strict compliance regimes to minimize mistakes in a mind-bogglingly complex system of signals intelligence collection. Imagine further that when the programs became public, there was a firestorm anyway. Imagine that nearly half of the House of Representatives, pretending it had no idea what you had been doing, voted to end key collection activity. Imagine that in response to the firestorm, the President of the United States—after initially defending the intelligence community—said that what was really needed was more transparency and described the debate as healthy. Imagine that journalists construed every fact they learned in light of the need to keep feeding at the trough of a source who had stolen a huge volume of highly classified materials and taken it to China and Russia.
What would you do? Here’s what: You’d take a hard look at your most forward-leaning programs—and you’d turn them off. You would do this using words like “prudential” and “current environment”—of course standing by the programs’ legality in some formal sense, just as the president has stood by you in some formal sense. But just as the president has let the intelligence community swing in the wind, limiting his own exposure by making the problem all your own, you would cut your losses. You wouldn’t even be wrong to do so.
And you would do it knowing somewhere in your heart that some day, the pendulum would swing the other way and there would be recriminations for turning those programs off, just as there are now recriminations for having such programs online. You would even know that many of the same people would be responsible for the mutually contradictory recriminations. You would know that after some big attack or intelligence failure, the scoop that you turned off collection tailored to the sort of information you needed to stop that event would be just as irresistible to the Washington Post and the Guardian as was the story that you ran riot over Americans’ civil liberties. You would know that the papers would be just as careless with the facts. You would know that the same members of Congress who are today outraged at what your agency is doing would wax outraged then at what it isn’t doing. And you would know that almost nobody will bother to know what they are talking about before having very strong opinions about how you fell down on the job and thus bear responsibility for both the smoldering ruins of some federal building somewhere and for destroying American values.
I wrote this, as I acknowledged at the time, without any evidence that the retrenchment I hypothesized was taking place. I could feel it, but I didn't know it for sure.
Now I do.
In the two months that have followed this post, many signs---some of them public---have confirmed my sense that a significant curtailment of SIGINT collection is taking place. The most public sign is the White House's announcement in USA Today that things are under review: "the president has directed us to review our surveillance capabilities, including with respect to our foreign partners. We want to ensure we are collecting information because we need it and not just because we can," writes the president's counterterrorism adviser, Lisa Monaco. Privately, people have been more direct. A lot of activity that used to be happening is no longer happening. Other than a certain confidence that we are no longer spying on the leaders of allied countries in the way we were only a few months ago, I still don't know the parameters of the retrenchment. What I am sure of is that it amounts to a significant rollback---not in law but in policy---of certain activities that only recently were assumed to be solidly in bounds.
I have a bit of a pit in my stomach about this. I suspect we're going to regret it. But I also have pit in a my stomach on another point: I worry that the retrenchment, however broad it is, is not well theorized. That is, I worry that we are disclosing a large volume of material and rolling back surveillance authorities without a particularly clear vision---at least not clear to me---of what sort of material we should be making public and what sort of lawful authorities we should be putting off limits. I suspect, rather, that the impulses driving these decisions are highly reactive to bad press and negative blowback from individual governments.
David Kris, in his recent paper on Section 215 collection, hinted at this problem (see pp. 64-65):
As of this writing, it is not clear whether the Obama Administration intends to pursue a narrow or broad re-calibration. In public, at least, it has not clearly described the philosophical approach underlying the disclosures it has made, the limits on such disclosures, or a comparison between the current attitude and historical standards—although such thinking may well exist behind the scenes. To some observers, however, we seem to be on course for an environment in which the basic existence of all (or most) signals intelligence programs is publicly disclosed, with information about particular participants in those programs (e.g., providers and targets) still secret. That would be a very significant re-calibration. Whether and to what extent the transparency would extend still further, to other intelligence programs—e.g., those involving Humint or covert action—also remains to be seen.
The effects of a broad re-calibration could be felt in at least two ways. First, official disclosures of previously classified information will resonate through FOIA and State Secrets doctrine, where the government’s litigating positions will be tested for consistency with the logic implicit in the voluntary transparency. It therefore may be difficult to predict exactly how such official disclosures may beget additional disclosures as compelled by the courts, especially in the absence of any overtly described philosophical approach.
Second, and perhaps more importantly, there is a potential interaction between increased transparency and the scope of intelligence activity. Intelligence activity that helps the U.S. government when done covertly may harm it when done overtly. For example, clandestine surveillance of foreign government officials may aid U.S. foreign policy—e.g., by giving U.S. treaty negotiators insight into their foreign counterparts’ instructions. As such, foreign policy makers may support and even require such surveillance from the Intelligence Community. On the other hand, however, transparent surveillance of foreign government officials may have precisely the opposite effect, creating challenges that cause policy makers to require less surveillance. If less surveillance leads to a perceived intelligence failure, of course, resulting demands to expand surveillance may cause the pendulum to swing back.
I'm not saying that no re-calibration is in order right now. To the contrary, if the intelligence community cannot keep its secrets about activity that causes enormous damage when it becomes public, some re-calibration is necessary---a fallback to positions that are more defensible when revealed. But it should be a re-callibration both as to disclosures and as to substance that is thought through at a philosophical level. What sort of activity are we going to refrain from---and are we going to refrain as well from recriminations when our restraint produces negative consequences for our interests? What sort of activity that we used to conduct in secret are we now going to conduct, at least in broad policy terms, in the open---and are we willing to have much-less-effective SIGINT in order to do our SIGINT in public? For that matter, why should not the same theory imply retrenchment in other intelligence spheres? Should we be disclosing---as Kris provocatively suggests---in broad policy terms our human intelligence capabalities and programs too and curtailing those HUMINT programs that would embarrass us if revealed? And if not, why not? Is there some theoretical difference between stealing secrets electronically and stealing secrets by human hands?
I cannot say that these questions are being elided in the current retrenchment. But nor have I seen much evidence that there is any serious or systematic consideration of them either. Rather, I suspect that the retrenchment is being driven by nothing more or less than reaction to the last leak---and fear of the next one.