What is the CIA For?

One of the many lasting legacies of Donald Trump will be the introduction of the term “Deep State” into the lexicon of American public life. Trump has accused this “Deep State”—basically meaning the U.S. intelligence apparatus—of “running amok” and manipulating American politics and the law against him, during his Presidency and since. The various investigations have disclosed mistakes and even some abuses by the intelligence agencies, but Trump’s warmed-over, 1970s-vintage Leftist hysteria about the Central Intelligence Agency (CIA) and the rest of the “intelligence community” being evil Nazis is no more convincing now than it was then.

There is an irony at work here. If there was one moment when Trump seriously differentiated himself from, and began pulling ahead of, the rest of the Republican field during the primary process, it was a tumultuous debate in February 2016, described at the time by one perceptive observer as being “like a horrible Thanksgiving where the family decides finally they’re all going to say what they really think about each other”. Trump found wedge issues where he stood alone against the other Republican contenders, and his stance had the larger market share with Republican voters. A key such issue was the invasion of Iraq, where Trump cast himself (falsely) as the only Republican candidate to oppose the invasion—and, this being Trump, he added actual praise for Saddam Husayn.

The background narratives in the minds of Americans that Trump was benefiting from when he made this pitch—the idea the invasion was based on policymakers misusing intelligence and other deliberate “lies”, with the implicit suggestion that poor Saddam was no threat at all and fell victim to a neocon frame-up—were narratives disseminated and reinforced by the CIA as part of its bureaucratic warfare to exculpate itself as the mistakes in the Agency’s pre-invasion intelligence assessments began to manifest in the summer of 2003.¹ In other words, Trump’s path to the Presidency involved exploiting a situation shaped by the CIA behaving towards President Bush in an insurrectionary manner that really was something like what the MAGAverse believes happened to their guy.

Trump and Iraq, political issues that will be debated to the end of time, are not our questions here. They will be discussed here only in so far as they help answer what is our question for today:

If the CIA can, especially over Iraq, show such remarkable skill in manipulating information when it comes to the domestic bureaucratic struggle, whether it is forwarding the policy positions the CIA is not supposed to have within the inter-agency process or controlling the public narrative and deflecting blame from itself onto assigned scapegoats in other parts of the government, why is the CIA’s record as a foreign intelligence agency—the thing it is paid for—so abysmal?

This can be answered by examining the CIA’s (official) functions—of which there are basically four, with considerable overlap—and having a look at the evolution of the Agency over the last seventy years.

CIA FUNCTION ONE: INTELLIGENCE COLLECTION

The CIA is supposed to collect intelligence abroad, partly through intercepting communications and primarily by recruiting foreign spies, to provide information that can be utilised by policymakers in formulating their options and decisions.² At this, the CIA performs poorly, and it performs poorest in those areas where policymakers’ concerns are highest: international terrorist groups and aggressive totalitarian regimes.

It is a simple matter of record that the CIA failed to infiltrate Al-Qaeda before 9/11, and from a most sympathetic source we know what the first words uttered by Director of Central Intelligence George Tenet were after the planes hit the towers: “I wonder if it has anything to do with this guy taking pilot training?” As well he might have. Tenet was one of the few people in the world to be in any position to know what Zacarias Moussaoui was up to when he was arrested a month earlier, and to have done nothing about it.

The Agency had also known for seventeen months that two proven associates of Usama bin Laden were within the United States, and nothing had been done about them, either. The CIA did not even tell the Federal Bureau of Investigation (FBI) about Nawaf al-Hazmi and Khalid al-Mihdhar, who had landed in California soon after attending the final planning meeting for the “Planes Operation” in Kuala Lumpur in January 2000. In August 2001, about a month before Al-Hazmi and Al-Mihdhar immolated themselves and nearly 200 others by smashing Flight 77 into the Pentagon, the FBI began a search for the pair, having found out about them quite by accident, because of a personnel transfer. Even at this point, the CIA was less-than-fully-cooperative because of the infamous “Wall”,³ erected by the Congressional reforms of the 1970s, that legally separated “domestic” intelligence collection, with its emphasis on criminal prosecutions, from “foreign” intelligence work in a wholly unrealistic manner.⁴ The lack of coordination between the CIA and FBI engendered by “The Wall” was the single greatest cause of U.S. intelligence failing to detect 9/11.⁵

Given the CIA’s operational practices it could not have gotten spies within Al-Qaeda, neither by turning one of its members, let alone getting one of its intelligence officers into the group. The CIA relies very heavily on intelligence officers under official cover at the Embassies and liaison relationships with foreign intelligence services. The CIA does have, within its Directorate of Operations (DO), a program for officers with “non-official cover” (NOC)—what the Russians call “Illegals”⁶—but the cover stories (or “legends”) they tend to operate under, namely businessmen, professionals of some kind (often scientists), and (more in the beforetimes than at the present) journalists, do not lend themselves to infiltrating Islamic militant groups.

In an article in The Atlantic in 2001, Reuel Marc Gerecht, a former CIA case officer in the DO, wrote:

Intelligence collection on al-Qa’ida can’t be of much real value unless the agent network covers Peshawar. During a recent visit, at sunset, … I would walk through Afghan neighborhoods. Even in the darkness I had a case officer’s worst sensation—eyes following me everywhere. … No matter where I went, the feeling never left me. I couldn’t see how the CIA as it is today had any chance of running a successful counterterrorist operation against bin Ladin in Peshawar, the Dodge City of Central Asia.

Westerners cannot visit the cinder-block, mud-brick side of the Muslim world—whence bin Ladin’s foot soldiers mostly come—without announcing who they are. No case officer stationed in Pakistan can penetrate either the Afghan communities in Peshawar or the Northwest Frontier’s numerous religious schools, which feed manpower and ideas to bin Ladin and the Taliban, and seriously expect to gather useful information about radical Islamic terrorism—let alone recruit foreign agents.

Even a Muslim CIA officer with native-language abilities (and the Agency, according to several active-duty case officers, has very few operatives from Middle Eastern backgrounds) could do little more in this environment than a blond, blue-eyed all-American. … An officer who tries to go native, pretending to be a true-believing radical Muslim searching for brothers in the cause, will make a fool of himself quickly. …

The only effective way to run offensive counterterrorist operations against Islamic radicals in more or less hostile territory is with “non-official-cover” officers … But as of late 1999 no program to insert NOCs into an Islamic fundamentalist organization abroad had been implemented, according to one such officer who has served in the Middle East. “NOCs haven’t really changed at all since the Cold War,” he told me recently. “We’re still a group of fake businessmen who live in big houses overseas. We don’t go to mosques and pray.”

A former senior Near East Division operative says, “The CIA probably doesn’t have a single truly qualified Arabic-speaking officer of Middle Eastern background who can play a believable Muslim fundamentalist who would volunteer to spend years of his life with shitty food and no women in the mountains of Afghanistan. For Christ’s sake, most case officers live in the suburbs of Virginia. We don’t do that kind of thing.” A younger case officer boils the problem down even further: “Operations that include diarrhea as a way of life don’t happen.”

Behind-the-lines counterterrorism operations are just too dangerous for CIA officers to participate in directly. When I was in the Directorate of Operations, the Agency would deploy a small army of officers for a meeting with a possibly dangerous foreigner if he couldn't be met in the safety of a U.S. embassy or consulate. Officers still in the clandestine service say that the Agency's risk-averse, bureaucratic nature—which mirrors, of course, the growing physical risk-aversion of American society—has only gotten worse. …

The Directorate of Operations’ history of success has done little to prepare the CIA for its confrontation with radical Islamic terrorism. Perhaps the DO's most memorable victory was against militant Palestinian groups in the 1970s and 1980s. … Still, its “penetrations” of the PLO … were essentially emissaries from Yasir Arafat to the U.S. government. …

In my years inside the CIA, I never once heard case officers overseas or back at headquarters discuss the ABCs of a recruitment operation against any Middle Eastern target that took a case officer far off the diplomatic and business-conference circuits. Long-term seeding operations simply didn’t occur.

The problem of language capacity at the Agency would get some attention in the mid-2000s, but as Gerecht sketches here, that was an outcome, not a cause, of the fundamental problem: the CIA is simply too risk-averse to fulfil the collection mission against jihadists. In theory, Americans pay the CIA so that it will risk the lives of dozens of its officers to gather intelligence that saves the lives of hundreds or thousands of Americans. That the Agency takes a safety-first view is not, as Gerecht also notes, an Agency-specific phenomenon: this conduct is essentially by popular demand of the society, which is less willing to accept casualties of any kind than their forebears.

Much the same held true when dealing with the Islamic State (IS). That President Barack Obama’s decision to abandon Iraq at the end of 2011 was reckless and cynical does not really excuse the CIA for being, as even its most sympathetic chroniclers concede, blindsided by the IS rampage that created its “caliphate” in June 2014. The Agency worked for nearly a decade in Iraq alongside the U.S. military, and for more than a decade before that the country had been an intelligence priority: to have developed no effective stay-behind network after a quarter-century is a failure by any definition. Still, after a rocky start, the U.S.-led Coalition did begin rolling back the caliphate in late 2015. The major U.S. breakthroughs in gaining visibility inside the caliphate came about not due to “HUMINT” (human intelligence) infiltrations of IS, but generally because of “SIGINT” (signals intelligence), namely cracking IS’s online and other communications systems.⁷ This allowed the U.S. to launch targeted raids that killed or occasionally captured IS terrorists and scooped up the organisation’s internal documentation, beginning an iterative process that allowed the U.S. to map the IS network and take out its leadership.

As IS’s most senior leaders were protected by the capable amniyat (security service), with its Saddam-like compartmentalised and overlapping structure, these sectors had to be broken separately, which they occasionally were, again relying on SIGINT. Abdurrahman al-Qaduli (Abu Ali al-Anbari), the caliph’s deputy and almost certainly his designated successor, was destroyed in March 2016, and the evidence is that SIGINT was key.⁸ IS’s Media Department has long had a special role in the security of IS’s very top leaders and a breach in that communications system in the late summer of 2016 led to the elimination of media emir Wael al-Ta’i (Abu Muhammad al-Furqan) and the famous spokesman Taha Falaha (Abu Muhammad al-Adnani).⁹

There have been claims that Israel’s MOSSAD managed to gain access to IS’s Syrian “capital”, Raqqa, and Mohammed Emwazi (Abu Muharib al-Muhaji), known to the British tabloids as “Jihadi John”, was allegedly killed in that city in November 2015 after one of Emwazi’s disillusioned associates was recruited by the Secret Intelligence Service (SIS or MI6): once Emwazi’s location was passed to SIS, it was apparently forwarded to the cousins, and Emwazi was “evaporated” by an American Hellfire missile.¹⁰

There are indications, then, that Western intelligence agencies, managed to recruit a small number of agents within IS. This was not completely unprecedented from the earlier war with Al-Qaeda, but it does seem to have been on a relatively larger scale with IS. The intelligence from these infiltrations would have been shared with the CIA—by definition in the case of Five Eyes partner Britain—but there is no clear evidence the CIA recruited any unilaterals within IS, which would only restate the reliance on liaison relationships for HUMINT against “hard” targets. The media claims that the CIA had “developed sources in Syria and Iraq who are producing reliable information” raised more questions than they answered: a careful reading suggests the CIA’s HUMINT sources were in Syria and Iraq, but not in IS. One obvious source of intelligence was the Syrian rebel groups the CIA was ostensibly supporting in this period, who had a modestly successful record against IS. The reference could also be, for instance, to villagers the Agency had managed to contact in areas occupied by IS, or the CIA could have—as other anti-IS States did—recruited informants from among the smuggling networks that IS used to transfer its cadres across the Syria-Turkey border.

What is much clearer is that there was no significant effort for the CIA to infiltrate its officers into IS. There is no doubt this would have been easier than was the case with Al-Qaeda, since such officers could have used the cover of the huge flow of European and American foreign fighters joining IS, which included people with overt military backgrounds. Nor is there any doubt this would have been an extraordinarily dangerous mission and some of the officers would have been discovered and killed, something we can be reasonably sure—given the nature of the CIA and the U.S. more generally—we would have heard about by now.¹¹

The institutional process of neutering the CIA began half-a-century ago. The CIA, founded in 1947 with the express mission of preventing another Pearl Harbour, evolved out of the Office of Strategic Services (OSS), led by “Wild Bill” Donovan, which operated behind Nazi lines in parallel with the British to, as Winston Churchill demanded, “set Europe ablaze”. The OSS never really achieved much as an intelligence agency and one of its operations had to be stopped to prevent disrupting the ULTRA program that was the crowning achievement of Allied intelligence in the Second World War. Still, the OSS was brave and, up to the 1960s, some of this spirit remained as the CIA tried to repeat the program against the Soviet Empire, with “rollback” operations to free the Captive Nations and activities against Soviet-allied governments in what was then called “the Third World”. The results of the former were uniformly disastrous and the latter distinctly mixed, but the action-oriented culture kept the Agency trying. In the 1970s, as the Sixties Revolution began to bed down, the changing mores against the presumptions of authority and the need for secrecy in government led to President Richard Nixon’s downfall,¹² and one of the aftershocks of this was a sharp turn against the intelligence apparatus.

In 1975, the Church Committee, the Senate Select Committee of Investigation (SSCI) led by a Democratic Senator from Idaho, Frank Church, and its counterpart in the House of Representatives, the Pike Commission, led by Democratic Representative Otis Pike of New York, investigated ostensible abuses by the CIA, FBI, and NSA. As all these agencies had operated under Presidential orders, there was little that was actually illegal, but the liberals on the Congressional committees were out for blood, casting U.S. intelligence, the CIA above all, as a rogue actor, threatening civil liberties and freedom tout court, at home and abroad. The exposure of the CIA’s foreign actions was presented in the most stark moralistic terms, with much focus going on activities like assassinating (or attempting to) foreign leaders, and the Agency’s experimental programs, some of which were quite strange; they were also designed as red team thinking and many had already been closed down.¹³ The assistance rendered to the CIA by foreign governments that maintained public postures divergent from the U.S. on, say, Vietnam, notably Sweden, were exposed, and thereafter turned genuinely hostile for decades to recover credibility with their own populations.

The Church Committee presented U.S. intelligence as evil, and implicit in that was the claim the CIA and the others were competent, perhaps brilliant. Americans had become increasingly distrustful of all institutions after the 1960s and Watergate had hardly helped. This dynamic was especially acute on political Left, making it most receptive when the Soviet Union and its surrogates within the West amplified and exaggerated the findings of the Congressional Committees. This was the context in which the books under the name of Philip Agee were put out by the KGB and Cuba’s DGI, an active measure that publicly identified 2,000 or more CIA officers and agents around the world.¹⁴ Self-criticism is a special feature of the West, but unconstrained by perspective it becomes masochism—a personal vice that becomes mortally dangerous at a societal level. Congress aired the CIA’s activities in the 1970s; there was no parallel process in the Soviet Union. The KGB was able through terror to keep hidden the fact that it played a much larger role in world affairs than the CIA, both in terms of the KGB’s geographic reach and the scale of its subversive activities within individual countries, including in countries like Cuba and Chile that became bywords for CIA malfeasance. This gross disjuncture between reality, and the popular and scholarly understanding, of intelligence activities outlasted the Cold War. Whole histories are still being written, on Soviet foreign policy (which was ever-more tightly controlled by the KGB from the 1960s onwards) and of the “Third World” States where that great Cold War contest played out, that omit mention of KGB covert action entirely. The result is “the intelligence equivalent of the sound of one hand clapping”.¹⁵

Battered reputationally and demoralised in the 1970s, the CIA then had a series of reforms forced upon it that in a legal sense gave Congress oversight of the intelligence apparatus and in practical terms curbed the CIA’s appetite for forward-leaning activity. The concern at the Agency shifted from doing its nominal work of recruiting spies outside the United States and became covering itself at home, avoiding getting caught up in another Congressional witch-hunt. It was against this background the CIA was transformed into “a highly secretive and extraordinarily expensive Department of Motor Vehicles”. The CIA ceased to be a playground of adventurers and idealists, becoming dominated by careerists wanting stable employment and nature’s bureaucrats—people who want to keep things ticking over and avoid rocking the boat; bland, unimaginative, and often frankly mediocre. The shift in personnel at the CIA brought on by the Congressional reforms was reflected in the Agency as a whole.

The CIA, stripped of its “exceptionalism”, turned its focus away from a sense of mission against foreign adversaries to playing the game in D.C. as one government department among many, intensifying the “iron law of institutions”: maximising power within the institution (in this case, the U.S. government), as against maximising the power of the institution to fulfil its nominal purpose, with the corollary that that it is preferred that the institution “fail” at its stated purpose so long as power is retained within the institution (again, meaning the U.S. government here). What the Congressional critics of the CIA did not realise was that they had set the CIA up perfectly to play this game: the intrusion into the “intelligence community” opened up a two-way street. The Committees had tethered the CIA to the political system, and in so doing made it a primary concern of the Agency’s to influence developments in the elected sectors of the U.S. government, while giving them the capacity to do so. Fears about the CIA’s role within the United States was a potent motivation for the 1970s reforms; the effect of the reforms was to make that problem far worse.

This is the background leading up to 2001, when Gerecht described the essentially mythical CIA counter-terrorism program—and the reasons why the CIA would not even try to improve on the point. Which it would not have, absent 9/11. It was the attempt beginning in September 2001 to put together an actual counter-terrorism program, incidentally, with virtually no Arabic-speakers or trained interrogators, that explains the “enhanced interrogation techniques” and other methods the CIA adopted to try to obtain information quickly from and about Al-Qaeda: though presented now as a moral parable of “torture” and evil—a narrative given an official stamp in 2014 by Democrats on the SSCI, a re-run of the Church Committee—the real story is an ad hoc, incompetent, and mismanaged effort to make up for lost time.¹⁶

Gerecht’s article, published a few weeks before 9/11, is hauntingly prescient on a number of points. For instance, Gerecht documented the particular problem for the U.S. of having no NOCs and relying on the “host” in Pakistan, the Inter-Services Intelligence (ISI) agency: “American intelligence has not gained and will not gain Pakistan’s assistance in its pursuit of bin Ladin.” That was to say the least of it. During NATO’s time in Afghanistan, the insurgency was often described as having three major components—the Taliban, Al-Qaeda, and the Haqqani Network—but these operationally indistinguishable elements formed a single fluid network under the control of the ISI, and, of course, the Pakistani military-intelligence establishment directly sheltered Bin Laden on its own territory, a stone’s throw from the Pakistani Military Academy. Gerecht went on to note that destroying the Taliban regime by fracturing the tribal and other social foundations it rests on would be more effective than any clandestine operation in debilitating Al-Qaeda, and that the ideal figure for such a mission was right there, Ahmad Shah Masud, the last man standing of the anti-Taliban Mujahideen, the leader of the “Northern Alliance”. Yet no CIA officer had even spoken with Masud before October 1999. Bin Laden evidently agreed with Gerecht: two days before the 9/11 massacre that Bin Laden must have known would draw the U.S. into Afghanistan against him, Al-Qaeda’s leader had Masud murdered to remove from the chessboard the most effective asset the U.S. could have allied with.

When it comes to States, the CIA did have some success in the long struggle with the Soviet Union. The CIA’s first major spy in the Soviet system, Pyotr Popov, an officer of military intelligence (GRU), was recruited in Vienna in 1953 and worked as a valuable in-place agent for five years. It is true that Popov was a walk-in—someone who volunteered his services (usually by walking into a U.S. Embassy), rather than being cultivated by the CIA—but it is also true this happened in the context of the CIA’s Operation REDCAP, launched in 1951 to encourage walk-ins and defectors from the rezidentura at the Embassies, from KGB officers in diplomatic delegations abroad, and from personnel deployed at the Soviet military bases in the “Satellite States” in the East. Another REDCAP success was Pyotr Deriabin, a KGB officer, also based in Vienna, who defected in 1954 and provided important information about the internal workings of Soviet intelligence.

The Popov story did not have a happy ending, however: he was detected and shot in 1960. One possibility is that this was due to a breach in British intelligence: George Blake. The CIA’s next spy in GRU, Oleg Penkovsky, recruited in 1960, provided important information to the U.S. during the Cuban missile crisis—though to what effect and with what intent is debated. That very month, October 1962, Penkovsky was arrested; he was put to death in May 1963. This suggested a second possibility: there was a mole at the CIA. As explained below, that possibility haunted the Agency for a decade. Then there is GRU Major General Dmitri Polyakov, a most murky case: apparently recruited in 1965, recalled to Moscow in 1980, and executed in 1988, there is a grave suspicion Polyakov at least started out as a dangle, i.e. he remained under Soviet control and was used to feed disinformation into the U.S. system. The length of time between the recall and ostensible execution is also deeply suspect: Did Polyakov die peacefully and Moscow announced the execution to further the ruse? Did the KGB take another look at the files and decide Polyakov had over-shared? We don’t know.

It is fair to say that Polyakov is more representative than the two Pyotrs when it comes to the CIA’s recruits in the Soviet Empire. Vitaly Yurchenko, a senior KGB officer who ostensibly defected to the U.S. in August 1985 and then went back to the Soviet Union three months later after reconsidering, was almost certainly a dangle from start to finish. The complication is that Yurchenko did give up two real KGB agents, Ronald Pelton (at NSA) and Edward Lee Howard (at CIA), and one can imagine Moscow Centre claiming it awarded Yurchenko the Order of the Red Star for his “infiltration operation” against the CIA as a way of saving face; stranger things have happened. It is also possible Yurchenko gave up Howard—who was transferred safely to the Soviet Union—to sow confusion to protect Aldrich Ames, the CIA counter-intelligence officer the KGB had then-recently recruited who was already showing his worth. For a time, the CIA did believe Howard was the source of the compromise they knew had occurred, which turned out to be Ames.

It might seem implausible that the Soviets would expose their own agents, but in fact this is standard tradecraft for Moscow. The KGB (and its successors in modern Russia) are prepared—as no Western service is—to give up vast amounts of true information and genuine agents if it will protect a more important interest, either an ongoing disinformation campaign or a better-placed agent. This consideration is what has caused controversy over Adolf Tolkachev, an electronics technician working on Soviet radar, among other things, who was signed up by the CIA in 1979, apparently arrested in 1985, and shot in 1986, having been betrayed by Howard just before his defection and probably by Ames as well. But, there is the fact Tolkachev made a half-dozen approaches to U.S. diplomatic vehicles before he made contact with the CIA: Is it possible the KGB did not notice this? There are doubters. The sceptics’ case is that what Tolkachev provided, while genuine, was of little value to the U.S., and Tolkachev was part of an active measure to reinforce the CIA’s belief that the Soviet Union was technologically and economically powerful. In short, Tolkachev told a (carefully curated) truth in service of a bigger lie, which is always the best disinformation, and the KGB gained access to the technical equipment the CIA used to run spies in Moscow into the bargain. Maybe.

While the CIA’s record against the Soviets is mixed (let us be charitable), the record against Saddam’s Iraq is much clearer: the CIA had no assets worth the name to give the U.S. visibility into the workings of Saddam’s government and no unilateral spies at all on the two crucial files, the WMDs and terrorism. This did not stop the Agency producing analytical products that spoke with confidence on both issues, while concealing their threadbare evidentiary basis. The apparent recruitment by the CIA of a network of informants inside Iraq on the eve of the invasion led to the decision to deviate from the plan that had taken months to prepare and initiate hostilities with an attempted decapitation strike against Saddam. Tenet, in his blustering way, was confident his spies had located the tyrant at the Dora Farms in southern Baghdad, and President Bush felt an obligation to shorten what turned out to be a brief three-week campaign as far as possible to spare lives on all sides.¹⁷ But the airstrikes that began the Iraq invasion at about 05:30 local time on 20 March 2003 did not kill Saddam and it is not clear Saddam had ever been at the complex. Whether the CIA’s network was compromised by Saddam’s secret police or was simply wrong, it was an ominous foreshadowing of the intelligence failures that would be revealed in subsequent months and years.¹⁸

The CIA was unable even to acquire intelligence that would give it insight into Iraqi society, which could have been augmented by recruiting human sources but did not necessarily require them. The extent of the Islamization of Iraq under Saddam, for example, could have been detected by following the State media. The CIA actively ignored this, largely because acknowledging Saddam’s Islamist turn would have undermined the Agency’s theory that Saddam was a secularist who could not cooperate with religious fanatics, the foundational assumption of much of the CIA’s analysis over Iraq before the invasion, which distorted its assessments of Saddam’s relationship with Al-Qaeda. If people are looking for “politicised intelligence” over Iraq, this is it. The post-invasion investigations were unanimous: there was not one single case of policymakers improperly pressuring the CIA, though there was evidence that CIA analysis had been less-than-rigorous in its assessments. This is not an exception. The politicisation of intelligence almost always occurs on the intelligence side, with the hiding of evidence and the slanting of analysis that cannot be detected behind a classification wall, not on the policy side, where any attempt would be leaked to the press by the CIA in minutes. Amazingly, the CIA continued to claim vindication on its non-cooperation theory after the invasion, despite the theory being openly falsified as the Saddamist remnants combined with the jihadists to form the insurgency. Another factor here was the CIA’s scorn for open sources. For the Agency, information given by a foreigner willing to betray his country for money is, by definition, more credible than information available in the public domain—from academic outlets, journalistic inquiry, émigrés,¹⁹ and increasingly social media.

The CIA’s responded to its mistakes over Iraq—which had, among other things, conditioned the deployment of insufficient U.S. troops to maintain order—not with any contrition, but by waging a scorched earth media war against other elements of the bureaucracy that compounded the damage by destroying the administration’s strategic messaging capacity.

In the 1990s, the CIA’s networks in Clerical Iran were compromised by shoddy counter-intelligence (see below) and stripped bare. In 2002, with the “axis of evil” business and the public revelations of Iran’s clandestine nuclear-weapons program, the CIA was once again having to make up for lost time to keep up with policymaker demands for information on a serious national-security threat. In late 2009, the Islamic Republic was forced into further admissions about the extent of its nuclear program, and in 2010 a campaign of assassinations against the staff preparing these genocidal weapons began. It was clear Iran had been infiltrated, which was the news for the CIA. The other shoe dropped when the clerical regime began a hunt for the mole(s), in the course of which, as Yahoo News reported last year, Tehran discovered something even more valuable: the “internet-based covert communications system” the CIA was using to maintain contact with its assets inside Iran. This “elementary system”, as one CIA officer called it to Yahoo, was essentially a series of websites. It was never meant for extended use in hostile environments policed by capable intelligence services; it was a stop-gap system, originally developed on the battlefield, but it seemed to be working, and after a time the CIA concluded it had no need of the planned upgrade. Nothing is more permanent than the temporary.

Iranian counter-spies are hardly the most capable in the world, but once a breach has been highlighted they are quite effective at working it through. The Iranians turned one of the CIA agents and the CIA failed to recognise it was running a double agent. By the time the Israelis tipped the Agency off, the Iranians had broken into the communications system—using Google. From Yahoo: “once the Iranian double agent showed Iranian intelligence the website used to communicate with his or her CIA handlers, they began to scour the internet for websites with similar digital signifiers or components—eventually hitting on the right string of advanced search terms to locate other secret CIA websites. From there, Iranian intelligence tracked who was visiting these sites, and from where, and began to unravel the wider CIA network.” The Iranians picked apart the CIA network throughout 2010 and extinguished it by the middle of 2011. The CIA ceased all attempts to recruit agents inside Iran and was consumed with the full-time task of racing against the clock to rescue its exposed agents. And the disaster did not end there.

Iran is in a tight strategic relationship with Russia, and this axis also includes China. Tehran forwarded what it had found to Peking, and with this guidance “Chinese intelligence obtained physical access to the transitional, or temporary, secret communications system used by the CIA to correspond with new, unvetted sources — and broke through the firewall separating it from the main covert communications system, compromising the CIA’s entire asset network in that country”. In 2011-12, the CIA’s assets in China were liquidated entirely, with dozens of people being shot. The Russians were duly informed, and, though the damage to the CIA’s operations in Russia itself was contained (“Aspects of the CIA’s Russia operations have historically been walled off from the rest of the agency, which likely helped minimize the damage”), the CIA realised in 2013 it was in the midst of a global calamity. In the Middle East, the CIA’s communications system with its agents basically collapsed as the Iranians began applying what they had learned in their imperial provinces, in the Fertile Crescent and Yemen. Not a lot of spying got done as the CIA’s agents tried to figure out how much danger they were in—and if any spying was done, the Agency was unable to find out about it, since it could not reliably contact its agents.

This being the CIA, it turned out to be worse than it looked: the CIA had been warned as early as 2008 about the flaws in the communication system by John Reidy, a contractor employed “to identify, contact and manage human sources for the CIA in Iran”, and the CIA’s only response had been to fire Reidy. (The CIA’s use of contractors for such tasks is yet another highly dubious practice.) It goes without saying that nobody was fired at the CIA for causing a worldwide disaster that cost God-knows how much in lives and treasure. Remarking on the incentive structure this creates for malpractice and corruption, and what it says about the supposed Congressional oversight mechanism, one former official bitterly concluded to Yahoo, “our biggest insider threat is our own institution”.

The apparent anomaly with Russia in early 2022, where the CIA made a rare public display of its knowledge about Vladimir Putin’s intentions to conquer Ukraine, was at best the exception that proves the rule. On the one hand, the CIA’s access to Putin’s inner circle was not that great: tracking the movements of conventional armies is (in theory: see below) one of the easier tasks the CIA performs; the bulk of the meaningful intelligence came from signals intercepts; and, as a legacy of the Cold War, the Agency’s Russia-focused analytical cadre is unusually strong. (As mentioned above, Russia is in general a somewhat special case at the CIA.) On the other hand, to the extent the CIA did have human intelligence sources, the Russian system is more open than a truly “hard” autocracy like China or North Korea. The qualifier also has to be added that the CIA made serious mistakes even here: it believed the Russian army would roll over Ukraine within days, two weeks at the most. This drove President Joe Biden’s decision—after all the sound and fury of the pre-invasion intelligence-derived messaging campaign that was supposed to “deter” Moscow—to do so little in practical terms to help Ukraine resist the invasion: the U.S. had already moved to planning how to support the expected insurgency by the remnants of the Ukrainian government and army. The most that was done for Ukraine’s President Volodymyr Zelensky was to offer to evacuate him, an offer he famously, bravely refused. To say that everyone got this wrong is only to restate the problem: just as many academics and analysts shared in the CIA’s major mistake over Ukraine, it was possible for analysts with no access to classified information to share in the CIA’s greatest success, understanding months before February 2022 that the Russians were going to invade Ukraine and that the intention was to capture Kiev and destroy the State, not to expand the Russian holdings in the Donbas.

CIA FUNCTION TWO: COUNTER-INTELLIGENCE

The CIA is responsible for coordinating U.S. counter-intelligence, which is to say the prevention of enemy spies penetrating America and, where that is not possible, the discovery and neutralisation of such spies. There is a systematic neglect in U.S. intelligence—not just at the CIA—of counter-intelligence, one of the most basic and the most important aspect of espionage, since failure at this task makes all the others impossible, as adumbrated above with the 2009-13 disaster in Iran and China. The CIA’s counter-intelligence record is atrocious and really always has been.

The OSS was riddled with Soviet spies from the time it was officially created in June 1942. Duncan Chaplin Lee, one of the senior assistants to the OSS chief, William Donovan, was a Soviet agent. OSS simply did not take the Soviet threat seriously because Donovan was so maniacally focused on the anti-Hitler war he naively believed Stalin and his Communists were allies.²⁰ The OSS had not one single agent in Moscow.²¹ In late 1944, the OSS finally caught on to the problem of Communist infiltration and began compiling lists of suspected traitors.²² The OSS was dissolved in 1945 and by the time the CIA was created in 1947 President Harry Truman’s loyalty-security programs were operational, allowing proper vetting of the OSS remnants moved into the new CIA, meaning that the Agency was not seriously compromised by the Soviets at its founding,²³ despite the reasonable suspicions of FBI chief J. Edgar Hoover and the Chairman of the Joint Chiefs of Staff, General Omar Bradley.²⁴ The insouciance of the OSS about counter-intelligence would, however, become a defining feature of the CIA.

The CIA itself might not have been compromised in its early years, as the West was recognising it had long been in a Cold War with the Soviet Union, but Britain’s intelligence system had been turned inside-out by the Soviets, and the wartime Anglosphere intelligence practices had forged the Five Eyes system, a nearly-unique example of a State alliance—between the U.S., Britain, Canada, Australia, and New Zealand—where the members collaborate practically seamlessly on intelligence and (more or less) do not spy on one-another. Kim Philby, possibly the most infamous Soviet spy and one of their most successful ever recruits, had been head of the SIS counter-intelligence division focused on the Soviets from 1944 to 1947, a crucial reason Britain was much less successful than the U.S. in combing Communists out of its intelligence apparatus after the war,²⁵ and in October 1949 Philby arrived in Washington as the SIS liaison to the CIA.

At the time Philby took up his post in D.C., the CIA, still infused with the OSS spirit and many of the operatives who had carried out the guerrilla warfare in Nazi-occupied Europe, was engaged in “rollback” operations, directly challenging the Soviet conquest in Eastern Europe, most prominently in Albania. Philby was able to betray these operations down to giving the KGB’s predecessor the coordinates where the CIA officers and agents would parachute in. Hundreds of men disappeared behind the Iron Curtain, all of them shot,²⁶ frequently after gruesome sessions in Soviet “interrogation” dungeons, once so well described as a “gallery of fanatics and alcoholics in a chamber of horrors”.²⁷ Covert action operations continued to be blown, the high-level U.S. efforts to infiltrate the Soviet Union continued to fail, and the red flags mounted, but the CIA never became suspicious of Philby. It was only the defection to Moscow of Donald Maclean and Guy Burgess, two of the “Magnificent Five” (the Cambridge spies), in May 1951, that finally tipped the CIA off about Philby, who was effectively expelled from the U.S. and resigned from SIS in July 1951.

If CIA counter-intelligence foul-ups abroad were perennial, the question of penetrations within the CIA per se from the 1950s to the 1970s takes us into some very murky waters indeed. The question essentially hinges on an assessment of the CIA counter-intelligence chief from 1953 to 1973, James Jesus Angleton, who was—depending on which faction of the “intelligence community” and/or which historian one asks—a misunderstood genius or a crazy conspiracy theorist who crippled the CIA by pursuing a lunatic witch-hunt against an imaginary fifth column.

Angleton had taken up his post just as Britain was digging through the rubble to assess the damage the “Magnificent Five” had done and a decade later, in the early 1960s, Angleton became convinced there was a similar problem at the CIA. Angleton’s mole-hunt split the IC very badly, causing serious dysfunction at the time and lasting acrimony afterwards. The saga that consumed the latter half of Angleton’s career started in December 1961 with an apparent defection from the KGB: Major Anatoliy Golitsyn walked into the Helsinki Embassy, and thereafter provided a lot of information that turned out to be true, some historical (e.g., about the Five) and some current (e.g., about ongoing breaches in British intelligence). Golitsyn also provided information that was falsified, notably that the British Prime Minister Harold Wilson was a Soviet agent, the Sino-Soviet split was a ruse, and the “Prague Spring” was a KGB concoction.²⁸ Most importantly, Golitsyn said the KGB had a senior agent at the CIA, and Golitsyn said that Moscow Centre would soon send a plant to discredit him so that the Agency would ignore the mole.

In February 1964, the U.S. got another apparent KGB defector, Yuri Nosenko, who had been working for the CIA since June 1962 and claimed his position had become untenable. Nosenko began giving information to the CIA that conflicted (sometimes quite specifically) with parts of what Golitsyn had said. Angleton was not alone in thinking Nosenko was the plant that Golitsyn had foretold. Nosenko was kept in captivity until 1967 and subjected to harsh interrogation methods, one of the last times before 2001 the CIA was seriously involved in interrogation.²⁹ Ultimately, in 1969, Nosenko was officially ruled to be a genuine defector and was hailed as a hero at the CIA until he died in 2008. The fallout of that verdict basically destroyed Angleton’s career—and his mole-hunt. Angleton’s position at the CIA was first downgraded and then terminated, years before the Church Committee, which would surely have finished Angleton off, began its work.³⁰

The controversy over Golitsyn and Nosenko lasts to this day. The KGB began running active measures against Golitsyn almost as soon as he defected—fabricating documents “showing” he was a criminal smuggler, for example—and in 1967 it tried to assassinate him. In combination with the records we now have of the KGB damage assessment, it is overwhelmingly probable that Golitsyn was a true defector. The question then becomes whether Golitsyn did more harm than good to the CIA by “infecting a small but troublesome [and powerful] minority of CIA officers with his own paranoid tendencies”, as one scholar put it.³¹ To judge that, one must come to a view about Nosenko. Nosenko’s first case officer, Tennent “Pete” Bagley, having initially believed Nosenko was genuine, quickly changed his mind, which derailed Bagley’s career alongside Angleton’s. Bagley, though, remained convinced to his dying day Nosenko was a plant and made a persuasive case to that effect.³² It should be said, there is a damage assessment for Nosenko in the KGB archives that is positively venomous in its description of him; there is no reason to think this document—produced for internal use, with no expectation it would ever been seen by foreigners—is not genuine.³³

As to the mole at the CIA, there are as many theories as there are authors. Many believe there was no mole, and that if anything it was Angleton who fell prey to Soviet machinations. Among those who believe there was a mole, the candidates are legion. A popular if unconvincing theory is that the traitor was William Colby, the new CIA Director in 1973 who closed ranks at the Agency around the message that there were no moles at the CIA and fired Angleton.³⁴

Until and unless another Vasili Mitrokhin comes our way from Russia, the case of Angleton’s mole will remain lost in the wilderness of mirrors. What is much clearer is that Angleton’s downfall devastated the counter-intelligence capacity of the CIA. The removal of Angleton and his acolytes eliminated human capital from the counter-intelligence division and it was never replaced: in the more careerist and domestic bureaucracy-focused atmosphere at the CIA after the 1970s reforms, counter-intelligence—the perennial bearers of tumultuous bad news, whose daily life is a conspiratorial wasteland of suspicion about their colleagues—became a highly unattractive employment path.

It was in this post-Angleton landscape that the CIA suffered potentially its worst compromise of the Cold War, in the mid-1980s. Occurring so late in the Cold War, this is a telling piece of evidence that without Mikhail Gorbachev’s “New Thinking”, which curbed the role of the KGB in Soviet society and foreign policy, the Soviet Empire might still be a going concern.

Aldrich Ames, the head of the CIA’s Soviet counter-intelligence division, was recruited by the Soviets in 1985 and only arrested in February 1994, after the long Cold War and the Soviet Union itself were gone. Murmurs about problems at the CIA in early 1985 died down after Edward Lee Howard’s defection, which was assumed to be the explanation, but in 1986 the Agency realised that something was badly wrong. The CIA initially thought the compromise was technical, either that the KGB had gained access to Agency communications or had placed a bug (or bugs) in strategic locations. By 1990, Ames had brought CIA operations against the Soviet Union to a virtual standstill: recruiting spies in Moscow had become impossible and the CIA was struggling to protect the assets it already had. Oleg Gordievsky, the KGB rezident (“station chief”, in American parlance) at the Embassy in London, had narrowly escaped execution when his work for SIS was exposed to Moscow Centre, probably by Ames (this remains contested).³⁵ 100 CIA operations and dozens of CIA agents in the Soviet Union were either endangered or killed because of Ames’ treachery. Ames came under suspicion as early as 1991, but the counter-intelligence investigation was mishandled. Among other things, a chummy atmosphere of the kind Angleton would never have countenanced surrounded Ames when he was subjected to a polygraph, ensuring the proper psychological pressures were not applied and allowing him to pass it. In 1992, the Agency finally put the pieces together; it was simply a matter of trying to catch Ames in the act after that.

Among the broad reasons why the U.S. settled on invasion as the means to depose Saddam Husayn was that everything else had been tried: massive internal rebellion, the most complete sanctions regime ever seen, and two CIA-backed coup attempts—all of which failed. The dire failure of the two coup attempts was largely due to poor counter-intelligence. As we will explore more fully in the next section, counter-intelligence problems have often derailed CIA covert operations.

In March 1995, Major General Wafiq Samaraii, Iraq’s military intelligence chief during Operation DESERT STORM who was purged soon after, tried to bring sections of the military to revolt against Saddam in Baghdad, coordinating his plan with the U.S.-backed Iraqi opposition—Ahmad Chalabi’s Iraqi National Congress (INC), Jalal Talabani’s Patriotic Union of Kurdistan (PUK), and Masud Barzani’s Kurdistan Democratic Party (KDP)—based in northern Iraq, and probably some factions of the Shi’i opposition in the south (it seems the Iranian Revolution’s Iraqi clone, the Supreme Council for the Islamic Revolution in Iraq (SCIRI), was at least aware of what Samaraii was up to). Both the Kurdish north and Shi’a south were protected by the Anglo-American no-fly zones at this time. The U.S. publicly denied all connection with Samaraii, but the CIA had been deeply involved in the preparations that relied on Republican Guard and Iraqi Army units defecting under pressure from the attacks in the north and south as Samaraii’s uprising took place at the centre. In the end, it was difficult even to call it a coup “attempt”: its operational security was dreadful, the KDP refused to act when it realised the operation was compromised and there would be no defections, and Saddam easily smothered it. Samaraii fled to Syria.

In June 1996, the CIA tried again. The Agency believed it had recruited several senior military officials in Baghdad and, working mostly with the PUK, there was an attempt to coordinate a Kurdish offensive from the north with a military rising in the capital. Chalabi told the Agency their operation was thoroughly infiltrated by Saddam’s intelligence services. Once again, operational security was laughable: the PUK communicated with rebel officers in Baghdad on open telephone lines. However, not even after Saddam’s secret police rang one of the CIA’s assets based in a neighbouring State to tell him Baghdad knew what he was up to and would murder his family, did the CIA listen to Chalabi. The result was calamitous: the mutinous military officials in Baghdad who were not plants were rounded up and dozens of them were put to death after show trials; the turmoil among the opposition reignited the war between the KDP and the PUK; and, in late August 1996, with Talabani receiving support from Iran, Barzani did what had seemed unthinkable and invited Saddam into Kurdistan. (Chalabi had, in fact, warned the CIA what Barzani was going to do—and was ignored, again.) In short order, Saddam’s army had killed 100 INC members and put the rest to flight.³⁶ With Barzani installed in the Kurdish capital, Erbil, Saddam’s troops withdrew from the area covered by the no-fly zone, and the KDP ejected the PUK from its main city in eastern Kurdistan, Sulaymaniya, pushing the PUK into Iran. The CIA severed all ties with the INC in February 1997, claiming that the INC had failed in its mission to compose the differences between the Kurdish factions and shown it could not lead a broad-based opposition movement. Chalabi did not accept that this was the end, though: he went public about the CIA’s failures, he helped bring the KDP and PUK to a peace accord in September 1998, and led the lobbying of Congress to pass the Iraq Liberation Act in October 1998, making it official U.S. policy to overthrow Saddam’s government.³⁷

This is the backstory to the CIA’s furious enmity against Chalabi, which severely impacted the planning and execution of the invasion of Iraq. The CIA tailored intelligence findings in 2002-03, specifically on the Iraqi population’s attitude towards the “external” opposition, with the objective of ensuring Chalabi did not have a prominent role in the “New” Iraq, and for a time at least the Agency succeeded in this political goal. But this meant U.S. policymakers were being guided—on issues far larger than how to handle Chalabi—by intelligence “findings” that bore little resemblance to reality in Iraq, indeed that had no intention of representing the Iraqi reality accurately: shaping U.S. policymaker decisions on the narrow subject of Chalabi was the point. This complete inversion of the CIA’s role—misrepresenting the empirical evidence and pushing a policy outcome—had catastrophic consequences for post-Saddam Iraq.

In 2009, the CIA thought it had finally recruited an important agent within Al-Qaeda. Granted, the CIA had not actually recruited Humam Khalil Abu-Mulal al-Balawi; that had been left to the Jordanians. But it was a CIA tip that led to Al-Balawi, an internet jihadist propagandist using the kunya “Abu Dujjana al-Khorasani”, being found and detained in Jordan, so when the GID announced that it had turned Al-Balawi in prison and could use him to get to Al-Qaeda’s number two, Ayman al-Zawahiri, it was almost an Agency victory. After Al-Balawi was dispatched back to “Af-Pak”, he reported to Amman that he, as a medical doctor, had personally tended to Al-Zawahiri in Pakistan, and had further chances for access since Al-Zawahiri was ill. Short video clips from inside a jihadist camp were sent to Al-Balawi’s Jordanian case officer, Captain Sharif Ali bin Zeid, a cousin of King Abdullah’s, no less, and the detailed medical history of Dr. Z that Al-Balawi provided checked out, in so far as the CIA could check it. The CIA and GID wanted to debrief Al-Balawi in person before arranging how he would lead them to Al-Zawahiri. There was some debate about where the meeting should take place—Al-Balawi wanted it to be in Pakistan—but ultimately it was decided to meet in Afghanistan, at Forward Operating Base Chapman in Khost, in the east, Al-Qaeda’s old heartland along the Durand Line.

Sixteen people—Zeid, multiple CIA officers and contractors, and an Afghan working for the CIA—were stood around Al-Balawi’s car when it entered Camp Chapman, after being waived through three checkpoints, on 30 December 2009.³⁸ One of the CIA officers even brought a birthday cake (Al-Balawai turned 36 on 25 December). Al-Balawi stepped out of the car, and as some of the assembled crowd moved towards Al-Balawi to search him—something one might have expected to be done at one of the checkpoints outside the base—he detonated the suicide vest he was wearing under his clothes. Five CIA officers, two CIA contractors, the Afghan security guard, and Zeid were murdered. Six other Americans were injured. The failure to adhere to the most basic counter-intelligence protocols brought about one of the darkest days the Agency has ever had. The only incident where more CIA officers were killed in one go was Iran’s blowing up of the Marine Barracks in Beirut in October 1983. As well as being a historic counter-intelligence debacle, the Camp Chapman bombing was a vivid example of the problems Gerecht identified before the War on Terror even began, in particular the overreliance on liaison relationships and the specific issue of Pakistan. The FOB Chapman fiasco was a microcosm of the fundamental policy tension, highlighted so many times and always ignored, which undid the West’s Afghan campaign: fighting an ostensibly variegated jihadist insurgency in Afghanistan that was in reality controlled by Pakistan, while maintaining Pakistan as an official ally. Al-Qaeda’s Al-Zawahiri had worked with the Haqqani Network, a formal component of the Taliban whose senior leaders are also Al-Qaeda members, to orchestrate the attack, which was overseen, facilitated, and paid for by Pakistan’s ISI.

Counter-intelligence investigations, by their nature, are usually secretive affairs, but they do occasionally become public and intersect with American politics, producing convulsive effects. The two primary examples are the attempt, in the late 1940s and 1950s, to repair the horrendous damage done to U.S. institutions by Soviet espionage under the cover of the “Grand Alliance” against Hitler, and the more recent effort to untangle Donald Trump’s relationship with the Russian government.

Leave the “McCarthyism” episode for another day. Let us focus on the Trump case. Without relitigating the whole saga, one hopes that there can be analytic agreement on two points. First, Trump has—on this subject, as on most others—nobody to blame but himself: his behaviour invited suspicion and U.S. intelligence would have been derelict in its duties if it did not investigate what was going on.³⁹ Second, during the investigation, U.S. intelligence made grave errors, both in the substantive sense of their conduct and in terms of optics, which is no trivial matter, given the security agencies’ need for broad popular trust to function in democratic societies.

The aspect that is most germane for our purposes is the “dossier”—the collection of intelligence reports ostensibly from sources in and around the Kremlin showing Trump was a witting collaborator with Putin’s Russia—compiled by former British SIS officer Christopher Steele. The dossier is not—as a simple look at the timeline shows—the key to all mythologies some on either side of the partisan spectrum think it is.⁴⁰ Nonetheless, “the Steel dossier” was a major factor in convulsing American intelligence and the American elite, despite the fact an averagely informed person, with no security clearance, should have needed one read-through of Steele’s reports to conclude that: (a) the non-boilerplate dossier claims were blatantly factually false; and (b) almost certainly laced with Russian disinformation.⁴¹ And that was before we knew that Steele’s project originated as opposition research for the Hillary Clinton campaign,⁴² or had any visibility into how the dossier was put together. The selling point of the dossier was Steele’s much vaunted reputation as a spy working against Russia, but even on its own terms this was irrelevant to the dossier’s credibility.⁴³ Steele used sub-sources within Russia,⁴⁴ and and his other sources—often relaying second- and third-hand claims—were revealed to be an aberrant parade of the congenitally shady and/or Democratic partisans with deep personal, financial, ideological, and other connections to Russian intelligence and its cut-outs.⁴⁵ Once Steele’s methodology was clear—i.e., sometime in 2017—that should have been the end of it. Yet, on it dragged.

For many years, former Western intelligence officials, some of them recently retired, notably Obama’s highly political CIA Director John Brennan,⁴⁶ and others working professionally on intelligence matters outside the government—journalists, academics, think tankers—would appear in the media and when asked a question about claims in the Steele dossier or Trump-Russia “collusion” more broadly, respond with words to the effect of: “We have lots of information showing Trump is in deep with the Russians—but we can’t talk about that; it’s secret”. Some surely believed it; some cynically saw the personal rewards on offer for saying such things. In either case, it created a highly asymmetrical information landscape, where such claims had the prestige of ostensibly being backed up by classified information—assumed by many (often wrongly) to be inherently superior to open-source—while the views of critics could be dismissed as simple ignorance since they lacked access to the relevant facts, and there was no way around this dynamic until the underlying evidence started to become public.

By late 2021, under pressure from the serial revelations that Steele’s employers and sources and collaborators all had ties to Moscow, Steele himself conceded that he might have been deceived by the Russians. Beyond some “#Resistance” dead-enders, few would contest this now.⁴⁷ But when it actually mattered, the U.S. intelligence system considered the possibility Russian intelligence had polluted Steele’s reporting only to reject it, and substantially relied on dossier reporting to illegitimately secure surveillance warrants against a Trump associate.⁴⁸ The temptation to give then-President Trump credit for denouncing Steele’s reports as “fake news” as soon as they were made public in January 2017 must be tempered by the fact that he always says that; the law of averages says he had to be right eventually.

The upshot of U.S. intelligence so badly mishandling the Steele dossier is that it injected the Russian disinformation therein into U.S. government institutions, further polarised and otherwise destabilised American society—a primary goal of the Russian government—and added to the negative partisanship that might be enough to put Trump in the White House again next year, an outcome the Kremlin has reason to hope for. The mishandling of the dossier has played into the Trumpian agitation that has discredited counter-intelligence work in general on the American Right for years to come (“Russia, Russia, Russia” is now a Trump laugh-line to dismiss any concerns about Russian espionage), making life much easier for Russia’s spies. And then there is the original counter-intelligence issue of the Trump-Russia relationship, now permanently obscured by the dossier. It is an interesting academic question which one(s) of these were the intent of the Russian Chekist(s) who conceived the “dossier” provocation.

Where defensive counter-intelligence is a shambles, the CIA shows little sign of even thinking about offensive counter-intelligence. If defensive counter-intelligence can be summed up as “detection”, offensive counter-intelligence is about “deception”: it is about hindering the functioning of the enemy’s intelligence apparatus. Crucial parts of this include turning enemy agents within your system into double-agents and otherwise feeding false information into the enemy system, encouraging the belief in things you want believed (negative things about allies,⁴⁹ the presence of non-existent moles to consume resources searching for them, politically destabilising lies (yes, like the Steele dossier)) and deflecting attention from things you wish to remain hidden (weaknesses in your own State, the existence of your spies in their system). The strategic aim of offensive counter-intelligence is not the random sabotage of the enemy’s intelligence capacity, but manipulating the information available so enemy policymakers make decisions of their own volition that you want them to make. The ultimate prize of offensive counter-intelligence is the capture of the enemy’s counter-intelligence system, thereby granting you control over their whole intelligence apparatus: with full visibility of the enemy’s knowledge, intentions, and capacities, you can cripple entirely their ability to wage the spy-war against you, and gain effective control of the enemy’s decision-making process, tailoring the intelligence available to policymakers and the analytical evaluations of the consequences of various options, thus significantly removing the element of chance in trying to get the enemy to take the decisions most advantageous to you.

The most successful State-to-State example of offensive counter-intelligence is what the Soviets did through the Magnificent Five to Britain, and what the Soviets did to Australia in the 1970s—that by extension compromised the whole Five Eyes system—isn’t far behind. What Britain did to the Provisional Irish Republican Army (IRA), rendering it operationally incapable by the 1990s, is a case in point against “non-State actors”. The Security Service (MI5) had always performed well in the spy war against the Provisional IRA and from 1980 the deputy head of the IRA’s counter-intelligence apparatus, formally known as the “Internal Security Unit” (ISU) and otherwise known as the “Nutting Squad”, was an MI5 agent. (It is possible the head of the ISU was also an MI5 agent from the mid-1980s; that is contested.) Britain was able to prevent the discovery of its agents, rescue those who were discovered, ensure that other agents passed screening when they were considered for promotions, and in general thwart the IRA at every turn. The Soviets were a dab hand with aggressive counter-intelligence operations against non-State enemies, too, using such tactics as central tools in the Cheka’s relentless war in Europe in the 1920s-30s against the “White” émigrés—the remnants of the Volunteer Army that had brought the Soviet regime as close to collapse as it would ever get in 1918-19—and in the “pacification” campaign against the nationalist insurgencies in the “Western Borderlands” (Ukraine and the Baltics) from 1944 until Stalin’s death in 1953. Tito’s Jugoslavija would take up these tactics when dealing with its own enemy émigrés or “sixth column”—the (mostly Croatian) nationalist opponents of the Communist regime based in Europe, America, and Australia—from the 1960s to 1980s.

CIA FUNCTION THREE: COVERT ACTION

Having collected intelligence about a situation, the CIA can be tasked with conducting covert action to influence the outcome. Sometimes this involves the CIA using indirect or “soft” means: seeking, for example, to ensure the promotion (or demotion) of a foreign official or to manipulate an antagonistic government into taking a decision that is advantageous to America (the overlap with offensive counter-intelligence will be noticed here). And sometimes this involves acting directly or using “hard” power, up to and including the assassination of foreign leaders and terrorists. The most widely known failures of the CIA usually relate to this aspect of its mission. Even among people otherwise sympathetic to the CIA, the Agency’s lamentable record at covert action became the subject of humorous aphorisms.⁵⁰ The CIA’s first Director, Admiral Roscoe Hillenkoetter, would presumably feel vindicated: he had objected to the CIA having a covert action mission.

As mentioned above, early in the period when the West recognised it was in a Cold War with the Soviet Union, the CIA and SIS attempted OSS-style behind-the-lines operations to free the Captive Nations, starting with Albania: Operation VALUABLE (1949-53) was a disaster that saw 300 men shot. At least in that case the Americans could blame the British: the counter-intelligence compromise was Kim Philby in London.

The CIA’s performance in the Korean War was dismal: at a strategic level, it missed the initial Communist invasion and misinterpreted the signals before Red China’s direct intervention, and its tactical-level performance as a paramilitary adjunct to the military on the Peninsula was not much better. But the war led to Americans accepting that in this new world, facing down the Communists’ sleepless attempt to rule the world, a higher level of military-security spending and a more active role in the world would be necessary for the foreseeable future. In selecting the tools for waging this long twilight war, the new President, Dwight Eisenhower, had a taste for covert action and the CIA’s budget and prestige was raised accordingly. The U.S. was still working in part with the Mother Country when it ventured onto this path, but it was becoming ever-more-obviously the dominant partner..

In the first two major covert operations to eliminate Communist-trending governments, in Iran in 1953 and Guatemala in 1954, the CIA got lucky: its operations per se failed, but the CIA was pushing with the grain of local elites alarmed at the direction their country was being taken in, and these elites capitalised on the situation to take matters into their own hands. The problem was that it was not only angry Sixties Leftists who believed the CIA played the decisive role in these events. The CIA believed its own myth, and fed it to Eisenhower, who thought he had proof of concept.

When the CIA tried to extend the Guatemala model to Cuba, after the Soviets colonised the island through Fidel Castro in January 1959, it led to calamity. The Soviets had learned from Guatemala, too, and moved quickly to secure Cuba as a revolutionary base in the Western Hemisphere. The haphazard CIA planning that had succeeded in Guatemala, against a partially-formed Communist government with a military still capable of independent action, did not cut it in Cuba, where the KGB had seen to it that Castro no longer had independent power-centres to deal with and had at his disposal a well-trained secret police force capable of surveilling and controlling the society, as well as State institutions. The CIA’s assumption that the Bay of Pigs landing in April 1961 would trigger a Guatemala-style internal collapse of Castro’s regime was swiftly falsified, and the second Agency assumption on which the operation hinged—that in the case of difficulties the U.S. would intervene directly to ensure Castro’s demise—was likewise soon falsified. Freshly sworn-in President John F. Kennedy was willing to allow Operation ZAPATA that Eisenhower had left to him to go forward, but he had decided in advance U.S. air support and Marines would be no factor in its outcome.

Counter-intelligence failures have frequently been the undoing of CIA covert action programs, from Albania in the 1940s-50s right through to the above-mentioned coup attempts in Iraq in the 1990s. Cuba was the worst case. With U.S.-abetted counter-revolution from the outside off the table, the CIA leaned further into the more targeted covert operations to bring down Castro that had begun before the Bay of Pigs, namely assassinating the Maximum Leader. Castro claims the CIA tried to assassinate him 634 times: specificity is the key with any lie and when it’s a whopper it has to be very specific.⁵¹ Undoubtedly, the Agency did try to help Castro join the choir invisible many times, and some of these plots might well have been quite outlandish, though the most famous ones—like the exploding cigar, the wetsuit contaminated with bio-weapons, and luring the diving enthusiast Castro to a garishly painted, booby-trapped seashell—were never actually tried, either being abandoned early or never getting past the concept stage. The needless public revelation of these schemes was part of the Church Committee’s vindictive desire to embarrass the CIA to weaken the political opposition to the Congress arrogating to itself an oversight role for the Agency.

The CIA being unable to get rid of Fidel Castro is an eternal humiliation, of course, and the practical consequences were worse: from this bridgehead of the Soviet Revolution, carnage was spread in Latin America right down to the end of the Cold War. The cause was the CIA being quite simply outclassed by Cuban intelligence—a problem that did not end with the Cold War. Only in the late 1980s did the Agency get a genuine Cuban defector, Florentino Aspillaga Lombard—a walk-in, incidentally—who revealed that all prior agents the Agency thought it had, four-dozen of them over thirty years, were dangles. (As alluded to above, the same turned out to be true—albeit nowhere else at such an extreme level—in the CIA’s operations in most other Communist States.) Few have managed a counter-intelligence debacle on quite this scale. Particularly upsetting for Langley was confirmation that Rolando Cubela Secades, the CIA’s prize agent in the 1960s, had been filleting the CIA’s networks in Cuba all the time the Agency believed Cubela was preparing for a coup against Castro.⁵²

Perhaps the only rival to Fidel Castro for the affections of the global Left during the Cold War was Salvador Allende, a long-term Soviet asset, who became president of Chile in an election that was manipulated—possibly decisively—by the KGB in 1970. Allende’s downfall and death in a military coup led by General Augusto Pinochet three years later was and remains for many Leftists a paradigmatic example of American evil: the CIA is held to have engineered the coup against a democratic socialist because of exaggerated fears about Communism and/or as an act of “economic imperialism”, meant to protect the interests of International Telephone and Telegraph (ITT). (The latter is a riff on the morality story of the United Fruit Company making Eisenhower get rid of Jacobo Arbenz in Guatemala in 1954.)

In reality, Chile is another example of failed CIA covert action. The CIA had been tasked with preventing Allende taking power in 1970, and failed at every stage: in the election, in trying to block the certification of Allende’s victory by parliament (“Track I”), and in trying to foster a military coup (“Track II”). The U.S. realised ahead of time that its recruits were incapable of Track II and called off the operation, but the CIA’s control of its assets was so ineffective that they went ahead anyway. The 1970 coup attempt—carried out against the express wishes of the U.S.—not only failed, but, in causing the death of the head of the Chilean Army, produced political effects that backfired spectacularly, rallying a nation that had been wary of Allende’s radicalism and visible connections to Castro and the Soviets around the necessity of following the constitution, ensuring Allende took office and discrediting Right-wing schemes to topple him.

What revived the conspiracies against Allende was his attempt to Sovietize Chile, ruining its economy and imposing levels of repression through a Soviet-trained and -directed secret police force that were deeply shocking in a country with a long constitutional tradition. By August 1973, the other two branches of the State—the courts and parliament—had declared Allende an outlaw, and, with the country in chaos, everyone was waiting for the military to step in. Every child in Chile knew a coup was coming and the CIA picked up on this RUMINT, but it was taken as much by surprise as everyone else when it was Pinochet—heretofore considered an Allende loyalist—who initiated the coup on 11 September 1973. There was no popular reaction to defend Allende and a lot of quiet relief. The Pinochet coup was a Chilean action taken for Chilean reasons, in response to conditions created by Chileans. As the Church Committee bluntly (if grudgingly) summarised: “CIA did not instigate the coup that ended Allende’s government”.

In the post-9/11 war against jihadism, the CIA’s emphasis shifted back towards its OSS paramilitary roots, and it has accomplishments to point to. The CIA’s leading role in the innovative invasion of Afghanistan to overthrow the Taliban-Qaeda regime stands as a shining success, notwithstanding the CIA’s Pakistani-influenced analytical failings that made it reticent about teaming up with the Northern Alliance,⁵³ and the Agency became proficient at killing terrorists, though, as mentioned, most of the targeting packets were formed using SIGINT and droning jihadists does have a way of reducing their utility as sources of intelligence. During Obama’s Presidency, a de facto take-no-prisoners policy was adopted because of the political position Obama got himself into over Guantanamo Bay, which radically reduced the U.S.’s understanding of the workings of Al-Qaeda and, notoriously, IS. The reliance on technical intelligence and technology has some serious downsides: the accidental killing of ten Afghan civilians in the drone strike following the Islamic State’s massacre at the Kabul Airport during the retreat from Afghanistan in August 2021 was because a propane tank was mistaken for a bomb, and there was no human source on the ground to tell the U.S. who was in the car. When the CIA did attempt traditional spy agency activities during the War on Terror, there were some spectacular mishaps (see Camp Chapman above).

Even in paramilitary terms, the CIA is unlikely to be able—and is unlikely to try—to do much in re-Talibanized Afghanistan. To launch a covert raid or a drone strike, the CIA needs to gather intelligence from within the country, and there is no sign it has any such capacity. (The killing of Ayman al-Zawahiri was in every respect a one-off.) All the old problems, including the CIA’s risk-aversion and reliance on a liaison relationship with an openly treacherous Pakistan, remain in place and militate against the Agency making an aggressive move over Afghanistan. There is also a powerful new factor pushing in the same direction: the political incentive for the Agency to collaborate in President Biden’s presentation of his abandonment of Afghanistan as other than a catastrophe. Devoting serious intelligence resources to Afghanistan would itself be an admission that Biden was lying when he said Al-Qaeda “would not be there”, and if any significant CIA effort was made to infiltrate the Taliban-Qaeda network it would surely incur casualties, which would become public and scupper Biden’s program of turning the page on Afghanistan.

CIA FUNCTION FOUR: ANALYSIS

The CIA is charged with analysing the intelligence it collects: this involves evaluating the credibility of sources, ensuring information streams are not contaminated by double agents or other counter-intelligence problems, weighing the value of one-off but perhaps well-placed reports by inter alia checking how it lines up with longer-term reporting (an especially delicate task when dealing with a one-off report where the costs are high in the case of error), assessing the confidence with which the evidence allows judgments to be made, and in general assembling the data points into something useful for policymakers trying to understand a problem. On the analysis side, the Agency’s record is absolutely calamitous.

The CIA’s inability to understand the Iranian Revolution at every stage and missing the collapse of the Soviet Union are clearly the most monumental analytical failures of recent decades. The memoirs of the Iranian Revolution are mutually hostile within and between every faction in the U.S. government and the Shah’s Court, but the one point they all agree on is that the CIA was worse than useless during the crisis. “The Soviet Union is a despotism that works”, said Deputy CIA Director Robert Gates in 1986, and in 1987 the CIA put on record a formal estimate that GDP per capita was higher in East Germany than West Germany. Later revisionist attempts to claim the CIA was on top of things during the Soviet collapse simply evaporate when set against the testimony of everybody involved that the Agency was clueless.⁵⁴

Not preventing 9/11 and believing Saddam had an active nuclear-weapons program and vast stockpiles of biological and chemical weapons of mass destruction join this roster of high-profile mistakes. The CIA’s record on nuclear questions in general is woeful: the Soviets, Red China, India, and North Korea all conducted tests more quickly than the Agency believed possible.

Detecting the complex societal forces and contingent events that lead to an eruption like the 2010-11 “Arab spring” might be more forgivable, but with the budget and technology of the CIA detecting large-scale military operations should not be that difficult, and yet the CIA has missed: the June 1950 Soviet/DPRK invasion of South Korea, Red China’s intervention on the Peninsula later that year, the 1973 Arab war on Israel, the 1979 Soviet invasion of Afghanistan, the 1990 Iraqi invasion of Kuwait, the 2014 Russian invasion of Ukraine, and the 2015 Russian intervention in Syria. The CIA did detect the Soviet build-up in preparation for the crushing of Czechoslovakia in 1968, but, without visibility into the Soviet system, the CIA was unsure of Moscow’s intentions and in its reports to the President the Agency hedged its bets: the intelligence reporting had the practical effect of encouraging the Johnson administration’s wish-thinking that the Soviets would pull back from the brink.

Part of the problem is that CIA analysts regularly have little information to work with, but the Agency’s incentive to preserve its bureaucratic standing means it will not admit when it does not know, so it has to make sweeping judgments from fragmentary evidence. In bridging the gap—assembling the scraps into a full mosaic—CIA analysis “has a way of confidently repackaging establishment biases”, as one former officer puts it, and other former intelligence officials make the same point about the tendency of Langley’s analytical shop, the Directorate of Analysis (DA), to reflect “mainstream opinion”. This is unsurprising, since, unlike DO field officers who tend to be cynical about ideology of all sorts, there is a strong liberal orientation in the DA. Worse, the CIA regularly slants the analysis it shows policymakers, arranging what little information it does have to support CIA theories and/or policy preferences, either by disparaging evidence that goes against the CIA’s theories and preferred policies, or more usually by hiding such evidence from policymakers altogether. As a matter of principle and law, the CIA is not supposed to do this, but what is striking is that from a utilitarian perspective the policies pushed by the CIA are almost unanimously deleterious to U.S. national security.

Just one example: the CIA’s 2007 NIE claimed that “Iran halted the [nuclear-weapons] program in 2003”, and that this “suggests [Iran] is less determined to develop nuclear weapons than we have been judging since 2005”. The headlines were written to suggest there was no nuclear-weapons threat from Iran, and the Agency made sure to tell anyone who listened that this was its view. The effect was to shape the political environment such that not only was a military strike on Iran’s enrichment facilities removed from the table: President Bush’s diplomatic push to isolate and sanction the Iranian theocracy was crippled. This was despite the fact that what a close reading of the CIA’s own evidence in the NIE actually disclosed was that the clerical regime had miniaturised aspects of the nuclear program and at most paused, not “halted” or stopped, the overt weaponisation work, probably in response to the Anglo-American invasion of Iraq. In other words, the CIA had yet again made a sweeping conclusion from fragmentary evidence, had done so for domestic political reasons, had forced its policy preferences on the government via an information operation to shape public opinion—and in this case, had, as a by-product, given cover to the lawless imperial and nuclear ambitions of a hostile Islamic Revolution that was at that very moment killing American servicemen.

CONCLUSION

In response to a presentation of the kind above, the CIA will say: “You hear only about our failures; we cannot talk about our successes”.⁵⁵ The difficulty is that for this to be true, when set against this record of known failure on every front,⁵⁶ the successes would have to be monumental. There may be moments where the CIA has been the victim, rather than the perpetrator, of asymmetric media coverage, but they are moments. Whatever else might be said of the CIA, it tends its public image superbly, even in real time, as we saw over Ukraine last year. It is simply not credible to believe the CIA is sitting on some stash of triumphs that would have to be so spectacular they changed the way we thought about history.⁵⁷

Is there, then, a way to make the CIA functional as a foreign intelligence agency? Daniel Patrick Moynihan, the former U.S. ambassador to the United Nations (1975-76) and thereafter a Democratic Senator (1977-2001), thought not.

Moynihan, operating from open-source information—a knowledge of the Communist milieu in New York, where he could see a faith in decline by the mid-1970s, and demographic studies of the Soviet Empire—concluded by 1979 that the Soviet Union was not long for this world. This brought him into conflict with the CIA, which believed the Soviet Union was moving from strength to strength. After the CIA missed the revolutions of 1989 that released the Captive Nations from a half-century of Soviet occupation, Moynihan proposed that the CIA be abolished, and its functions absorbed by the State Department. Moynihan put forward Congressional bills to this effect in January 1991 and January 1995. Moynihan’s view was considered deeply eccentric—neither of his bills were even voted on—and his specific proposal, for transferring CIA functions to the State Department, was unworkable and unwise.

Still, Moynihan had hit upon some of the structural problems documented in this essay. For instance, Moynihan recognised the problem with the CIA “culture of secrecy”, which creates an in-group/out-group dynamic that gives precedence to information simply because it is classified and disparages open-source information.⁵⁸ The fallacy behind believing on principle something said by a paid foreign traitor over publicly available and testable information should be obvious, and too often classification functions to protect shoddy analysis and the assumptions underlying it from outside scrutiny, instead of protecting sources and methods. To say that Moynihan undersold the damage done to the CIA by the post-Watergate reforms is in some ways to make his main point for him: that the CIA is beyond repair and has been for some time, which is why all the attempted reforms have failed.

It seems unlikely there is a means of terminating the Sisyphean struggle to reform the CIA, however. Even if one takes a Moynihan-type view that the only real solution is to abolish the CIA and start again from scratch, political reality is never going to allow that. All that can be done is to identify the most dysfunctional aspects of the CIA’s institutional culture—the risk-aversion, the lack of seriousness about counter-intelligence, the analytical habit of claiming more certainty than the evidence allows—and try to realign the incentives to promote changes.⁵⁹

The Steele “dossier” section of the article has been updated

---

REFERENCES

1. The two most memorable CIA failures over Iraq were the Agency’s belief Saddam had vast stockpiles of chemical and biological weapons of mass destruction (WMD) and was actively advancing his nuclear-weapons program. The post-invasion findings were that Saddam retained a latent capacity on all fronts that was in certain respects more dangerous than anyone in the Bush administration had known to accuse Saddam of. That evidence was only gathered because of the invasion, which could be claimed as a benefit of the policy, as could the fact the invasion ended the destabilising geopolitical uncertainty by certifiably disarming Iraq, but, of course, the certainty disclosed that the danger was less imminent than had been believed and, therefore, partly undermined the administration’s rationale for the invasion, especially because of the woeful messaging strategy.

2. Most signals intelligence (SIGINT) is handled by the National Security Agency (NSA).

3. The 9/11 Commission Report (2004), pp. 269-72. Available here.

4. One of the most shocking—and needless—revelations by the Congressional investigations of the 1970s was that the CIA had assassinated (or, more usually, attempted to assassinate) enemies of the United States abroad. This led to President Gerald Ford signing Executive Order 11905 in February 1976 that made various structural reforms to the “intelligence community” and outright banned “political assassination”, a prohibition that was tightened by Presidents Jimmy Carter (E.O. 12036, January 1978) and Ronald Reagan (E.O. 12333, December 1981). This created a serious legal problem for President Bill Clinton when it became clear, after Al-Qaeda’s August 1998 East African Embassy bombings, that Bin Laden would have to go. In a wonderfully Clintonian sequence of events, the President approved the killing of Bin Laden provided it took place in the context of a capture mission, then in December 1998 explicitly authorised the payment of Afghan anti-Qaeda tribal elements if they captured or killed Bin Laden, reasoning that killing Bin Laden was an act of self-defence instead of an assassination, and then in February 1999, when the CIA asked for the same authorisation that had been given to the Afghan tribesmen, Clinton signed-off on a Memorandum of Notification that returned to the August 1998 language of official approval only for capturing Bin Laden. The result was that President Clinton subjectively “wanted [Bin Laden] dead” and believed he had authorised covert action to make it happen, while the CIA never understood this to be the case and lamented how much easier their job would be regarding Al-Qaeda if they were permitted to kill Bin Laden. See: The 9/11 Commission Report, pp. 131-33.

5. The specific law underlying “The Wall” was the Foreign Intelligence Surveillance Act (FISA) of 1978. See: The 9/11 Commission Report, pp. 78-80, 399-421. There are accusations, based on very fragmentary evidence, that the CIA had kept watch on Al-Hazmi and Al-Mihdhar, and had kept the FBI out of the loop because the CIA was attempting to recruit one or both of them—which would have been illegal, as well as obviously a disastrous failure.

6. While CIA NOCs and Soviet/Russian Illegals nominally perform the same task, there is not really a comparison. The Illegals—in their professionalism, institutional prestige, and numbers—are in a different class to even the best NOCs.

7. The CIA’s involvement against IS after 2014 much more resembled the Joint Special Operations Command (JSOC) counter-insurgency tactics in Iraq a decade earlier—an iterative process of putting together a picture of the IS movement’s networks from captured operatives and documents, assisted by cyber-infiltration of the group’s communications, allowing the terrorist nodes to be eliminated seriatim. A classic early case was the first IS spokesman, the young man we know only as Abu Maysara al-Iraqi, who run rings around the Coalition for three years until a covert operation covering Baghdad’s internet cafés led to his downfall in June 2006. This is not a coincidence: the CIA’s Special Activities Division developed a seamless relationship with JSOC and NSA during that earlier struggle, and the template was readily revived.

8. There were media claims after Al-Qaduli’s demise that the CIA had “developed sources in Syria and Iraq who are producing reliable information” and that these human sources, in combination with the interception of IS communications, were behind the “increased success targeting [IS’s] leadership”. It seeming staggeringly unlikely that the CIA had recruited a HUMINT source close to Al-Qaduli, one of the most elusive and capable operatives IS has ever had, a man who managed to deceive the Americans even when he was in their custody. And parsing the wording, it will be noted that the actual claim was not that these sources were within IS. Subsequent revelations confirmed this hunch: Al-Qaduli had been located by communications surveillance of somebody he had contact with.

9. It was a breach in IS movement’s Media Department that led to the death of its founder, Ahmad al-Khalayleh, the Jordanian jihadist universally known as Abu Musab al-Zarqawi. (A week before Zarqawi was killed, the first IS spokesman, Abu Maysara al-Iraqi, was eliminated.) IS put in place further security measures to isolate the caliph’s personal security even more within the system after that.

10. Michael Smith (2019), The Anatomy of a Spy: A History of Espionage and Betrayal, chapter seven.

11. The loss of CIA agents in Iran, China, and elsewhere in the same period the “caliphate” was at its height have since been publicly reported and admitted by the Agency.

12. Both of Nixon’s predecessors, John Kennedy and Lyndon Johnson, had wiretapped journalists to try to discover their sources in the government, for example, as well as political opponents—and even allies—just to keep themselves one step ahead. Nixon actually did less of this.

13. Probably the best-known of the CIA’s odd programs is Project MKUltra, which took on a life of its own on the Leftist (and these days Rightist) fringes as a “mind control program”, though it was more mundane than that: it involved a search for drugs and psychological techniques that would weaken the resistance of prisoners to interrogation. It was also discontinued in 1973. While it is often said that one of the primary supporters of MKUltra was CIA counter-intelligence chief James Angleton, because he thought its techniques might finally unlock his mole hunt, the reality seems to be that this is one of those factoids that was first adduced based on a leap of logic that misread the evidence and then acquired the status of common knowledge by repetition.

14. Vasili Mitrokhin and Christopher Andrew (1999), The Sword and the Shield: The Mitrokhin Archive and the Secret History of the KGB, pp. 230-34.

15. Christopher Andrew (2018), The Secret World: A History of Intelligence, p. vii.

16. Ad hoc the CIA’s post-9/11 counter-terrorism program might have been, it was not starting completely from scratch. The lineaments were established during the Clinton administration, and carried out by many of the same people, including CIA Director George Tenet, a Clinton appointee (in 1996). The most obvious example is “extraordinary rendition: the apprehension—kidnapping, if we are to speak plainly about it—of terrorists and their transfer to the tender mercies of allied despotisms like Egypt’s. The first case was Tal’at Fu’ad Qasim, the leader of the “Islamic Group” (al-Jama’a al-Islamiyya). Qasim had fought the Red Army in Afghanistan and then got involved the Bosnian jihad, an episode often unjustly left out of accounts when telling the story of Al-Qaeda’s rise. Qasim was captured by the CIA in September 1995 in Croatia, which served as a jihadist logistics hub for foreign fighters and other resources flowing into Bosnia. Qasim was then handed over to the Egyptian government and secretly executed before the end of the year, per a death sentence that had been passed against him by a military tribunal in 1992. On the worst estimates of the treatment “rendered” captives received during the Bush administration, none of them were put to death, yet Qasim’s case languishes in obscurity in the West. It is impossible to avoid the conclusion that the partisan allegiances of most American journalists is the crucial variable here.

17. George W. Bush (2010), Decision Points, p. 254. The Iraqi death toll overall in the invasion phase was low, totalling (roughly) between 11,000 and 15,000, and of that number 3,000 to 4,000 were civilians. At worst, then, civilians were one-quarter of the fatalities during “Phase III”, a combatant to civilian ratio of 3:1. See: Carl Conetta, ‘The Wages of War: Iraqi Combatant and Noncombatant Fatalities in the 2003 Conflict’, Project on Defense Alternatives, 20 October 2003. Available here.

18. Donald Rumsfeld (2011), Known and Unknown: A Memoir, pp. 459-60.

19. Douglas Feith (2008), War and Decision: Inside the Pentagon at the Dawn of the War on Terrorism, p. 518.

20. The Sword and the Shield, p. 108. The paranoia in Moscow Centre led to the “Magnificent Five” being cut off twice—between February 1940 and December 1940/January 1941, and October 1943 and June 1944—on the basis that they were double-agents controlled by the British SIS and MI5. The “evidence” for this was inter alia that among the voluminous documentary material the Five passed to Moscow, there was no document showing the plan for a British invasion of the Soviet Union. Obviously, such a document did not exist because there was no such plan, but in Communist theology the “imperialist” powers were self-evidently always planning aggression against the Workers’ State, and the only reason such an important document was missing among the high-grade intelligence the Five were handing over was that they were part of an elaborate British deception operation. Documents stolen by Soviet agents in the OSS were likely important in convincing the Centre for the second and final time to exonerate the Five. See: The Sword and the Shield, pp. 85, 90, 119-21, 125.

21. The Sword and the Shield, p. 111.

22. The Sword and the Shield, p. 130.

23. Harvey Klehr, John Earl Haynes, and Alexander Vassiliev (2009), Spies: The Rise and Fall of the KGB in America, p. 329.

24. The Sword and the Shield, p. 143-44.

25. The Sword and the Shield, p. 125.

26. The Sword and the Shield, p. 155.

27. Robert Stephan (2004), Stalin’s Secret War: Soviet Counterintelligence Against the Nazis, 1941-1945, p. 72.

28. The Sword and the Shield, p. 184.

29. It has frequently been claimed, relying on Nosenko himself, that one of the CIA methods used to try to break him in interrogation was the administration of LSD, one of the substances the much-controverted MKUltra program had placed high hopes in. This transpires to be another myth surrounding Angleton: it was discussed at the CIA whether to administer LSD and “truth serum” to Nosenko, but Director of Central Intelligence Richard Helms vetoed the proposal.

30. Angleton was briefly, quietly hired back at CIA, but was gone for good by 1975.

31. The Sword and the Shield, pp. 184-85.

32. Before he died in 2014, Bagley wrote a book, Spy Wars: Moles, Mysteries, and Deadly Games, offering the definitive case that Nosenko was a Soviet plant. At a minimum, Bagley tears to pieces the story Nosenko told of his position at the KGB, what he had access to, and how he came to defect.

33. The Sword and the Shield, p. 185.

34. The Colby-as-traitor thesis was given a workout in novelistic form by William F. Buckley in Spytime: The Undoing of James Jesus Angleton.

35. Gordievsky, exfiltrated from the Soviet Union by British intelligence in July 1985, was an extremely brave man who had become a dedicated opponent of the Soviet system after the 1968 crushing of Czechoslovakia and an ideological devotee of the Western cause, demanding to be left in place as an agent rather than defecting so he could do maximal damage to the Soviet Empire.

36. ‘Report of the Select Committee on Intelligence on the Use by the Intelligence Community of Information Provided by the Iraqi National Congress’ (“SSCI Report on Chalabi’s INC”), 8 September 2006, pp. 23-24. Available here.

37. SSCI Report on Chalabi’s INC (2006), pp. 24-25.

38. Joby Warrick (2012), The Triple Agent, pp. 170-71.

39. The Senate investigations concluded in their analysis there had been no “collusion”—a word now imprinted forever more on the brain of every politically-conscious person in the Western world—between Trump and the Kremlin, but the exhaustive factual outlay in the Senate reports document the multiple points of contact between what we might call Trump World and Russian intelligence, notably Moscow’s ill-disguised WikiLeaks front-group, the conduit for some of the nastiest active measures during the 2016 Election. See: ‘Report of the Select Committee on Intelligence on Russian Active Measures Campaigns and Interference in the 2016 U.S. Election: Volume 5: Counterintelligence Threats and Vulnerabilities’, 18 August 2020. Available here.

40. Trump announced he was running for President on 15 June 2015, the (in)famous “golden escalator ride”. It was in the months after this, in late 2015, that U.S. intelligence began looking into the Trump entourage, after Britain’s Government Communications Headquarters (GCHQ) passed SIGINT to Washington showing “suspicious ‘interactions’ between figures connected to Trump and known or suspected Russian agents”. Before the year was out, another Five Eyes State, Australia, plus Estonia, Germany, and Poland, as well as possibly France and the Netherlands, had passed along similar material. Even in this post-9/11 situation, “the Wall” caused trouble: foreign-origin intelligence about a foreign counter-intelligence threat runs through the CIA, but the concern related to private American citizens, whom only the FBI can investigate. The CIA eventually took what it had to Congress in August 2016, by which time—beginning 31 July 2016—the FBI had opened an official counter-intelligence investigation it codenamed CROSSFIRE HURRICANE. Steele compiled his reports between April and December 2016, and the FBI first received parts of the “dossier” on 19 September 2016. On this last point, see: U.S. Justice Department Inspector General Report, ‘Review of Four FISA Applications and Other Aspects of the FBI’s Crossfire Hurricane Investigation’, 9 December 2019, p. 4. Available here.

41. There were a few notable exceptions, who correctly assessed Steele’s “dossier” from the get-go, including Soviet historian David Satter, former NSA counter-intelligencer and historian John Schindler, and the great historian of Soviet intelligence and intelligence generally Christopher Andrew.

42. The Clinton campaign having funded the production of “the Steele dossier” first became public in October 2017, and the deception surrounding this only became clear in March 2022 when the campaign and the Democratic National Committee were fined over $100,000 for concealing this opposition research project as “legal services” and “legal and compliance consulting”. The dossier advocates obfuscated the fact that Steele had worked within the framework of a partisan political operation by claiming that he had initially been hired by conservative anti-Trumpers at The Washington Free Beacon, funded by Republican donor Paul Singer. As early as February 2018 it was clear this was not true: the Free Beacon had contracted Fusion GPS from October 2015 to May 2016, primarily relating to Trump’s domestic business activities, and Clinton independently hired Fusion GPS in April 2016—a contract lasting until December 2016—to look at the Trump-Russia issue. There was, therefore, nothing gathered by the Free Beacon to be handed over to the Clinton/DNC operation, and Steele’s first memo or report in the “dossier” is dated 20 June 2016, a month after the Free Beacon contract with Fusion GPS lapsed.

But the narrative was persistent—it lingers even now—that Steele pursued the Trump-Russia story, as it were, independently, motivated solely by his concerns about the threat, and his alleged bipartisan funding to get the job done was held up as proof. The funding is not, in and of itself, the end of the debate over the dossier’s contents, but it is important context: there is no issue where knowing that the evidence is question is an opposition research project does not have some bearing on its credibility.

(It should be noted that Steele and his supporters even claim that Steele did not know the Clinton campaign was his employer through Fusion GPS until after July 2016. The purpose of this claim is to shield Steele from the charge he manipulated the U.S. intelligence system with political propaganda, since Steele first started sharing information with the FBI and Justice Department, in particular Bruce Ohr, that month. For an objective observer, Steele’s claim of ignorance about where his salary was coming from adds to the doubts about his honesty and/or his competence.)

43. The reputation of any ex-spy is based on their record, which is often by its nature difficult to check. Steele was said to have played a key role in breaking open several huge cases, specifically the Kremlin’s November 2006 assassination of Russian defector Alexander Litvinenko in London and helping the FBI uncover bribery at FIFA. As it happens, there is reason to doubt these claims: neither Litvinenko’s wife nor best friend, nor the former New York Times reporter Alan Cowell, who wrote a book about the Litvinenko case, had ever heard of Steele, and while the author of the most comprehensive book on the FIFA corruption scandal, BuzzFeed journalist Ken Bensinger, did know Steele, it was only because his encounter with Steele was memorable for how little Steele knew about the FIFA case.

Still, the truth or falsity of the claims underlying Steele’s reputation does not actually bear on the reliability of the dossier. Even if Steele’s stellar reputation is warranted, he did not collect the dossier reports directly; he received them at second-, third-, and fourth-hand. The credibility of hearsay at so many steps removed is obviously in doubt: even eyewitness testimony is often wrong, so accepting testimony about events from people who do not even claim to have seen it themselves is more fraught; stories often get corrupted (through no malice) by Chinese whispers; assessing the biases of sources—the political views and loyalties that impact what emphasis they put on a story when passing it along—is more difficult when the story has been through so many hands, especially if the story’s origins are murky; and so on. It is a matter of record with the Steele dossier that “sources” were sharing what they perceived to be unconfirmed rumours, plus information they had read in (or misread in or been told about from) the news media, and numerous “sources” were partisan Democratic operatives and people with connections to the Russian government.

44. Steele was widely acknowledged for the work he had done against Moscow in two decades at SIS/MI6, particularly as head of the Russia desk from 2006 until his retirement in 2009. For the Hillary Clinton campaign that hired Steele, his renown was the point: Steele’s name would be used to launder the opposition research that became “the dossier”. But there was the problem: Steele was so well-known—especially by Vladimir Putin’s government—that Steele himself definitely could not gather intelligence in Russia without being detected by the Kremlin. Steele, therefore, had to rely on “his sources in Russia”. Once Steele was running a primary sub-source, Igor Danchenko, who was drawing from a network of his own sub-sources, that meant, first, Steele’s own much-touted credibility was a useless barometer for the dossier (whatever one makes of Steele, that judgment has no bearing on sources at one and two steps removed whose claims form the actual content of the dossier), and, second, it was possible, if not probable, that Russian intelligence had coopted or coerced some of these sources into serving its purposes. The premier Western State intelligence agencies have difficulty operating in Russia without getting tangled in Moscow’s games; the idea a retired spy with a private network on a shoe-string budget could out-fox the FSB (and GRU) always seemed inherently implausible. It should be added that this is based on the assumption Russian intelligence was not in control from the start: Danchenko was the subject of an FBI counter-intelligence investigation in 2009-11 because he was suspected of being a Russian spy, and that investigation was terminated because the FBI erroneously believed he had fled the country, so there was never any clarity on what his relationship to Russia’s special services had been.

45. Igor Danchenko told the FBI he was Steele’s “primary sub-source” shortly after the dossier was made public by BuzzFeed in January 2017, and in July 2020 this became public knowledge. The revelation that the FBI had previously suspected Danchenko of being a Russian agent soon followed.

The November 2021 indictment of Danchenko for lying to the FBI revealed a lot of detail about how the dossier had been constructed. It was not flattering. Instead of information from well-placed sources in Russia, what Danchenko passed to Steele was musings and ramblings he picked up over drinks from a network of friends and social contacts—some in Russia, more in America. (As one observer well-summarised, “the Steele dossier—even though a former British intelligence agent wrote it—really was a product of the Washington swamp”.) In particular, Danchenko had drawn a number of “specific allegations” from Charles Dolan (“PR Executive-1”). Danchenko lying about being in contact with Dolan was part of the indictment. The other lie was claiming to receive a call from Sergei Millian, the head of the Russian-American Chamber of Commerce, alleging there was “a well-developed conspiracy of cooperation” between Trump and Putin. In reality, “Danchenko fabricated these facts”.

Dolan was not only a Democratic operative and Clinton ally; his professional milieu was connected strongly to the Russian State, including to “one of Danchenko’s Russian sub-sources”, namely Olga Galkina (“Sub-Source 1”), a Russian political official. Dolan had been working closely with the Russian Embassy in the U.S.—as had another sub-source, “Organizer-1”—to organise a conference in October 2016, and both went to Moscow in June 2016, where Danchenko already was. Russian diplomatic facilities, of course, are always staffed significantly by intelligence officers. Steele certainly knew this: he claimed in the weeks before the 2016 U.S. election that Russian intelligence officials had recruited a network to interfere in the process that was being run and paid “out of the Russian Consulate in Miami”. That there is no Russian Consulate in Miami was a sign of things to come.

Ironically, both Danchenko and Dolan are recorded in the indictment remarking on the other being uncomfortably close to the Russian government. Danchenko said of Dolan: “he’s a bit naïve in his … liking of Russia”. Dolan said of Danchenko: “I think he worked for FSB”.

Above all of these Russia-related problems with the creation of the dossier, there were the problems with the creators.

First, Christopher Steele. Weeks before Steele was hired by Fusion GPS to work on the dossier for the Clinton campaign, Steele had approached Fusion regarding work he was doing for Oleg Deripaska, one of Putin’s oligarchs, who was trying to recover money from Paul Manafort, a disreputable lobbyist who had briefly served as Trump’s campaign manager. (Fiona Hill, a Russia scholar and sometime Presidential Adviser, said she was “shocked” when she found out Steele was the author of the dossier because she had known him for years as someone “constantly try[ing] to drum up business” using his old contacts and she had no doubt Steele’s motives and approach provided Moscow the “perfect opportunity” to “feed [Steele] some kind of misinformation”.)

Second, Fusion GPS. In the spring of 2016—at the exact moment Steele was hired by Fusion to work on the dossier—the firm was working vigorously to help the Russian lawyer, Natalia Veselnitskaya, sent by Putin to the U.S. to try to block the passage of the Magnitsky Act. (Veselnitskaya has become famous as the Russian government representative at the June 2016 Trump Tower meeting with then-Candidate Trump, Donald Trump Jr., Jared Kushner, and Manafort.) Bill Browder, possibly the most prominent Russian oppositionist in the West, denounced Fusion GPS co-founder Glenn Simpson for working as an effective “agent for the Russian government, as an adjunct to the Russian FSB”.

46. Brennan was not inventing his centrality in this saga: he was the CIA Director when the Agency received the foreign SIGINT about Trump’s connections to the Russian government in late 2015, the beginning of the—quite reasonable—process of the U.S. “intelligence community” looking into the counter-intelligence risks posed by the Trump campaign. It was the way this fact was leveraged for political warfare purposes that was the problem. Brennan was ubiquitous on the op-ed pages and especially the television news networks, particularly on MSNBC, through 2019 arguing that Trump’s “collusion” with the Russians was a known fact, and spoke with the authority, always implicit and sometimes spelled out, of someone who had seen the knock-down proof in the classified files—the routine tactic of the CIA in managing the domestic media. (The CIA generally disseminates its narratives in the media either through on-the-record statements by “former” officials like Brennan, or via “leaks”—anonymous quotes to journalists from serving intelligence officers—but always with the approval of the Seventh Floor. This would be seen many times during the Trump administration, as it has with every other.)

47. The defence of the Steele dossier that remains—including from Steele—is that it was “raw intelligence”, which is to say a starting point for further investigation. This is in some ways an advance over the claims that the Steele reports were being steadily “corroborated”, but it does involve a new and rather audacious claim that, while the Steele dossier might be false in the specifics, it is true in general.

48. There is an argument—and not only from MAGA world—that Steele’s reports were such obvious junk the FBI must have been wilfully lying in using them as the basis for the FISA warrants against Carter Page. And it is pretty clear that the FBI used circular reporting—the “dossier” itself and media reports based on Steele’s “findings” as if they were separate streams of intelligence—in its presentations to the FISC. Allowances made for the fact that wilful deception and wish-thinking are not always as clear-cut in practice as is in theory, the “intelligence community” is still left with incompetence as the kindest explanation for its handling of Page and the FISA Court. The alternative explanation is deliberate violations of the law to pursue a political vendetta, and the evidence is not exactly 50/50 here. It always strained credulity that the FBI’s omission of Page’s past work as a CIA informant in the FISA applications was an innocent oversight, and subsequent investigation showed the FBI had wilfully hidden this information from the FISC.

49. In late 1950, Donald Maclean, one of the “Magnificent Five”, became head of the American desk at the British Foreign Office. This was during the most intense period of the Korean War. In addition to supplying the Soviets with vast quantities of classified documents, which Maclean’s deputy later concluded must have been “of inestimable value in advising the Chinese and the North Koreans on strategy and negotiating positions”, Maclean sought to foster Anglo-American tensions: “For perhaps the first time in his diplomatic career, Maclean showed open sympathy in a Foreign Office minute with the crude Stalinist analysis of the inherently aggressive designs of American finance capital. There was, he said, ‘some point’ to the argument that the American economy was now so geared to the military machine that all-out war might seem preferable to a recession produced by demobilization.” See: The Sword and the Shield, pp. 154-55.

50. Edward Luttwak once wrote: “With only a handful of exceptions, there have been only two kinds of CIA secret operations: the ones that are widely known to have failed—usually because of almost unbelievably crude errors—and the ones that are not yet widely known to have failed.” And there is a line from William F. Buckley—often said to have been in response to an effort to eliminate the Indonesian tyrant Sukarno (r. 1945-67), though the exact context and content are unclear (I can’t find the original source)—that an assassination operation had “all the hallmarks of a CIA operation: everyone in the room was killed except the intended target”.

51. The Church Committee—which it can be trusted did not under-count—said in its final report in 1976 that it had “found concrete evidence of at least eight plots involving the CIA to assassinate Fidel Castro from 1960 to 1965”. The Committee received a list from Castro of twenty-four U.S.-sponsored attempts on his life: it was concluded that all were fabricated.

52. Brian Latell (2012), Castro’s Secrets: The CIA and Cuba’s Intelligence Machine, chapter ten.

53. The CIA’s relationship with the Pakistani ISI created a very strange situation during the invasion of Afghanistan: the U.S. mission was to remove the Taliban-Qaeda regime, which was essentially a colonial administration imposed and sustained by Pakistan, specifically the ISI and the Pakistani Army, and yet the CIA was heavily influenced in its assessment of Afghan politics by the ISI. In the shadow of 9/11, the Pakistani military establishment understood it could not save the Taliban-Qaeda regime. That said, the uniformed cadre that owns Pakistan has always maintained an “office of hedging your bets”, so during the invasion the Pakistani military provided support to the Taliban-Qaeda jihadists, just in case it was possible that increased American casualties would force a retreat that allowed Pakistan to keep its occupation regime, and the ISI tried to condition the aftermath by shaping CIA analysis to be wary of the anti-Taliban Northern Alliance that formed the bulk of the ground troops working with the teams of CIA officers and U.S. Special Forces on horseback to coordinate military operations and call in airstrikes. The Pakistani “intelligence” fed to the CIA argued that allowing the Alliance, comprised mainly of Tajiks and Uzbeks, to capture the Afghan capital would inter alia risk civil war by alienating the Pashtuns, the group that dominates the south and east, along Pakistan’s border, and forms the backbone of the Taliban. The CIA was, as ever, not content to press this Pakistani-derived disinformation within the inter-agency process and leaked it to the press, stoking panic about a “quagmire” in a war that lasted all of six weeks. See: Known and Unknown, p. 376; and, War and Decision, pp. 78-79.

54. One of the most prominent early revisionist papers contains its own refutation: an interview with National Security Adviser Brent Scowcroft in which Scowcroft says the administration had no warning from the CIA that the Soviet demise was imminent—that is, not even at the end, after the August 1991 KGB putsch attempt against Mikhail Gorbachev. President George H.W. Bush has said the same. Colin Powell, Chairman of the Joint Chiefs of Staff at the time, said that during the Soviet endgame the CIA “could no longer anticipate events much better than a layman watching television”.

It can also be noted that the controversy over the CIA’s failure to detect the Soviet collapse displayed two other themes touched on in this essay, particularly in the context of Iraq after 2003.

First, major analytical failures are often not simply incompetence; all the “errors” run in one direction because they result from the CIA slanting the analysis with the intention of securing a policy outcome. Robert Gates, the Deputy CIA Director from 1986 to 1989 (and later Director), reported watching his boss in this period, William Casey, routinely “sit in meetings and present intelligence framed in terms of the policy he wanted pursued”.

Second, the CIA has been far more consistently adept at the Beltway game, played out in the media, of maintaining its bureaucratic-institutional standing in D.C., than it has as a foreign intelligence agency. In 1995, after a build-up of criticism over the CIA missing the greatest geopolitical event of the twentieth century, the Agency went on the offensive to claim it had seen the Soviet collapse coming by, among other things, circulating carefully selected morsels from intelligence reports where the Agency had alluded to weaknesses in the Soviet economy and had considered the possibility of a KGB coup against Gorbachev. As it happens, Bush Senior and Scowcroft state categorically that the coup took them by surprise. See: George H.W. Bush and Brent Scowcroft (1998), A World Transformed, pp. 520-21.

55. The quote is from William Webster, the CIA Director from 1987 to 1991. Something very similar was later said by the current CIA Director Bill Burns: “Our successes are often obscured, our failures are often painfully visible”. There certainly are distorted polemical attacks on the Agency—the most prominent example is Tim Weiner’s 2006 book, Legacy of Ashes: The History of the CIA (for a good review, see here)—and readers will have to decide if this article falls into that category.

56. Sometimes the CIA’s record will be defended by arguing that critics’ expectations are unreasonable—that “success” is defined as “predicting the future”, and this is a misunderstanding of what a foreign intelligence agency does or can do. This certainly does happen: Tim Weiner’s Legacy of Ashes is a case in point. I hope readers will bear me out as agreeing this essay has not done that. It has sought to examine the totality of the CIA’s mission: the collection of intelligence on targets of national-security concern to inform U.S. policymakers, influencing events abroad through covert action, counter-intelligence to protect the U.S. intelligence system and the country more broadly from foreign spies, and analysing collected intelligence to assess the nature and probability of threats. It is in this last category that the CIA has a predictive role, with its assessments informed by collected intelligence. If we take the collapse of the Soviet Union as an example. There surely are some who believe that the lack of a CIA report saying the Soviet Empire will be dissolved on 26 December 1991 represents “failure”. This is, of course, naïve in the extreme. The actual failure is that the CIA’s assessment of the trendline or “big picture” was that the Soviet Union was durable over the long-term.

57. To be clear: the finding would not be that the CIA had orchestrated some event or series of events. The highest success in the intelligence realm—as in many other areas of security policy, especially counter-terrorism—is when nothing happens. Having been so critical of the CIA, it is important to emphasise that this is done in the full knowledge the successes of an intelligence agency can be extremely difficult to see, not just in the literal sense that they are usually formally classified, but it involves a degree of imagination: the victory is that a State never went down the alternative historical road where some terrible disaster occurred. This is the realm of the counter-factual and those determined not to see what they were spared can do so.

58. ‘The Commission on the Intelligence Capabilities of the United States Regarding Weapons of Mass Destruction’ (“Iraq Intelligence Commission” or “Silberman-Robb Commission”), 31 March 2005, p. 174. Available here.

59. A lot of suggestions for reform of the CIA are to strip it of extraneous missions so it can solely focus on intelligence collection. This usually takes the form of advocating that the CIA hand over its paramilitary functions to the Pentagon, but detaching the analytical directorate, the DA, might be of more immediate value. Among other things, if the DA was made into an independent institution answering directly to the President, it would separate it from any group-think at the Agency—though it would still have its internal group-think issues, of course—and it would significantly diminish the interface of the CIA with the domestic political system. With some changes to the independent DA—for instance, slimming it down to include more analysts with deep expertise and fewer Bright Young Things—this arrangement could have the dual benefit of higher quality analysis of collected intelligence and a reduction in the CIA’s incentives (and mechanisms) to manipulate policymakers. The CIA would no longer be responsible for finished assessments, taking it away from the inter-agency decision-making process, and this would eliminate the Agency’s need to cover itself—through media leaks and the rest—if the assessments lead to unpopular decisions. The Agency would be judged, in the Beltway and among the population, by the quality of the intelligence it collected—incentivising it to succeed on that front.