During my career in a NATO country’s military establishment I had interactions with the Central Intelligence Agency often and some of the other US agencies occasionally. The characteristics of the CIA, as I saw them, are the subject of this essay. I think they are applicable across most of the forest of US intelligence structures.The CIA is a very large organisation and the result, bureaucratically speaking, of all these bodies and money is that it is very fragmented. Time and time again our guys would be talking to their guys about some country. Their political section would give us their views of who was in, who was out, power struggles and so on. We would then ask a question about how the economy fitted into this. A pause, well you’ll have to talk to our economy guy, and from the back of the room, blinking in the unaccustomed light, their economy guy would mutter something and then return into his dark burrow. Clearly, the political guys were the stars and economic guys were not: almost an afterthought. But in most countries the economy is the most important driver of politics most of the time; in some countries, it is almost the whole story. How can you possibly separate the two? I will confess, by the way, that we sometimes provoked this response for our amusement.

The politics people were completely obsessed with personalities. I will never forget the leader of their delegation in one of our meetings proudly handing out a piece of paper with the people around the Boss divided into several groups, each with a neat name: his Tribe, Security organisations, the previous Boss’s relicts I think they were. Imagine what that had cost the American taxpayer, the hours of discussion as to whether the Minister of Whatever was more of a Tribe than a Relict. I asked the CIA guy what was the point? What had we learned? How did this help us understand or predict? Of course it was all rubbish, the truth was that the Boss had put together a team; it was what that team did that was important, not from where he’d plucked its members. In another case we were all invited to make predictions about the future of a country. The CIA’s entry was a point series – if this guy becomes Boss, then this; if that guy gets in, then that; if somebody else, then something else. As it turned out, the new Boss wasn’t even one of the people on their list. But note their assumption that everything depended on who the new Boss would be. This obsession with personalities seems to be a built-in characteristic of American thinking for some reason and you see it in the political leadership, the media and intelligence all the time. Milosevic is the problem, Saddam Hussein is the problem, Ahmadinejad, Qaddafi, Assad, bin Laden, Putin. If we can only get this guy out, all will be well. No it won’t, all the objective local conditions that carried him to the top will elevate somebody similar. It’s not a person, it’s a whole country. But the Americans never learn, they force the Bad Guy out and they get either total chaos or a new Bad Guy who turns out to be rather like the old Bad Guy.

A third characteristic was illustrated by a presentation from the Defense Intelligence Agency. The author proudly presented a chart of combat missions by a certain country’s air force. The air force was said to be decaying with poorly maintained aircraft and untrained pilots. And yet his data showed numerous sorties in mountains in bad weather and no crashes. The obvious deduction was that the air force was in much better shape than we thought it was. But the author was so impressed with the data itself that this escaped him. And an impressive collection effort it would have been too, one that few intelligence structures could have carried out: that little chart had cost a great deal of money and involved some impressive and expensive technology. But I suspect that they are so often impressed by the collection technology that they forget why they are collecting it.

I mentioned these characteristics to a neighbour who had also had a lot of dealings with US intelligence and security structures. She agreed with my three and added a fourth: you never meet the same people twice. Indeed, I don’t recall ever meeting the same person twice either (and at her level and the nature of what she was doing – common security issues – the constant changing of the team would have had a much greater consequence). I don’t know how many people work in the US intelligence and security megalopolis but it surely is in the hundreds of thousands if not a million or two. A lot of turnover, cross-postings, promotions and so on. So maybe it’s not surprising that the teams do change quite a bit from year to year. Which raises the question, of course, of how much time on the job the average analyst has.

I passed this to a former colleague who has had much more experience with the CIA than I and he added yet another problem: It’s known in the military as SOPO – senior officer present’s opinion. He told me that in numerous private conversations CIA analysts had complained to him that their assessments were frequently sent back to them to be re-written to fit the conclusions the higher-ups had already come to. And that, of course, is fatal because it creates a closed loop in which you only hear what you already believe and worse, think it’s confirmed by the intelligence. I am always amused how much people are impressed by that phrase “confirmed by the intelligence”. If they only knew.

The purpose of intelligence is to minimise surprise and you can’t expect to do that if you compartmentalise things, obsess about personalities, get carried away by the collection mechanics, change your personnel all the time and confirm what you want to believe.

I’m not going to say, by the way, that my group got it all right: in this business 50% is a pretty good score and too many analyses fall back on the “maybe this, maybe that, time will tell” school of waffle. But we did try to look at the whole picture, regarded individuals as important but embedded in a context, didn’t have a lot of collection technology and therefore went more for what is now pompously called OSINT (open source intelligence) and didn’t do operations at all. And, thanks to one former boss who was still setting the style in my time, we were encouraged to stick our necks out. But we – and all other intelligence organisations – do suffer to some degree from SOPO.

Even without the bosses demanding the “correct” answer, getting it right is very difficult – imagine trying to solve a jigsaw puzzle when you don’t know how big it is, what the actual picture is and have a random selection of pieces some of which may be from another puzzle altogether. I used to amuse myself by asking old intelligence guys when in peacetime in the Twentieth Century had the intelligence guys got it right and convinced the politicians (not much point in the first if you can’t do the second). Personally, I can only think of Richard Sorge. Intelligence is much easier in wartime, by the way, because then you have a very good idea of what the picture is and how big the puzzle is.

Which, come to think of it, seems to be another defect of American intelligence – if you convinced that Russia is a permanent enemy, then everything it does will be interpreted either as an openly hostile act or hostile in some cunning, sneaky, back-handed, barely detectable way.

So, in my experience, the US intelligence structure has made a difficult job almost impossible with these self-imposed handicaps.