Monday, August 4, 2014

Two Views of Intelligence


This article originally appeared in Cicero Magazine on 19 June 2014.

The purpose of intelligence is to speak truth to power. Intelligence agencies exist to provide decision-makers with absolute, unbiased facts or, failing to obtain absolute facts, to provide as clear, unbiased and true a picture of a situation as possible using what facts are available to enable decision-makers to reach informed conclusions as to what course of action to take. Rarely is it the case that Intelligence can present a picture of a situation that is purely fact-based. There are common intelligence requirements, such as determining the intent of another party, which can never be known for certain because their very nature prevents certain, permanent determination. ‘Gaps’ in fact while attempting to form as clear, accurate and true a picture of a situation as possible can only be filled with conjecture or informed ‘guesswork’ based upon past actions, history, logic and/or statistical probability. Informed ‘guesswork’ is what intelligence analysts do.

One of the central arguments in the field of intelligence analysis regards which angle estimative or predictive strategic intelligence analysis should be approached. Two schools of thought have emerged: Straussians, based around University of Chicago political scientist Leo Strauss, and Kentians, around Sherman Kent, founding father of CIA’s estimative process.

Values or Truths?

The Straussian view of analysis is founded on the idea that the ‘regime’—specifically the form of government and society a state adopts—provides a window through which the political thought, intentions and actions of a state can be observed and predicted. It assumes there is a continual human search for which form of regime is ‘best’ and a qualitative analysis of the differences between these different forms is the route to determining which regime is ideal. For example, Strauss believed during the Cold War that the essential qualitative differences between American democracy and Soviet communism was the most important issue of the day.

At first read, there appears to be nothing controversial about that idea. However, the Straussian view requires that the judgment of the quality of a regime be based upon how well it provides public goods such as liberty, freedom, justice and so forth. This invites what can be called a ‘values’-based judgment into the process using determinants which are subjective in nature. This conflicts with the objective focus of mainstream notions of the social scientific approach which focusses on facts over values judgments and pictures measureable differences between regimes as a matter of different degrees of focus in pursuit of universal human pursuits.

The Kentian view of analysis is based firmly in the belief that intelligence analysis should be approached as an intellectual subject in the liberal social science tradition. Analysts were certainly not to make values judgments, but rather search for the underlying universal ‘truths’ common to man. Kent, a Yale historian, shied away from establishing or applying theoretical analysis based in international relations to intelligence, preferring instead more practical empirical frameworks and methods. He held that the intellectual and emotional detachment of his analysts allowed them to produce better estimates than military analysts or policymakers and their staffs because they focused on academic ‘truth’ as their goal without an attached or vested interest in their particular ‘regime’. Kent’s belief in the value of this objective analytical system was such that he held it to be more valuable than clandestine intelligence collection. No number of microphones or satellite photos could substitute for the value of being able to objectively divine the meaning of long-term trends in order to accurately estimate future actions.

While Kentian objective analysis can lead analysts down blind alleys due to strategic deception, Straussian analysis can lead one to look at regimes such as the Soviet Union or Saddam Hussein’s Iraq in light of worst-case scenarios because of their (supposed) diametrical opposition to the U.S. regime.

Kent created a system for intelligence analysis with the goal of creating ‘institutional memory’ so knowledge would become cumulative and not be lost between generations. While building CIA’s analysis structures Kent was very much committed to the professionalization of intelligence analysis. Besides its academic commitment to neutrality, the Kentian view was also shaped by organisational factors. The positioning of CIA’s early estimative body between the military and political leaders required it to maintain a reputation for objectivity so as not to become a target for either. Kent held up the CIA’s performance in more accurately estimating Soviet capabilities during the Bomber and Missile Gaps of the late 1950s compared with military intelligence estimates as an example of its detached view of analysis discouraging it from inflating estimates to support its own organisational goals. The Air Force was accused of inflating Soviet bomber and ICBM estimates in order to justify an ever greater share of the defence budget going to its SAC and U.S. ICBM programmes.

Leo Strauss never worked, studied or wrote about intelligence analysis. He was a political scientist. However, as Gary Schmitt and Abram Schulsky argue, Strauss’ work on political analysis can be related to intelligence analysis. Strauss argued that political and social sciences could never be true or ‘hard’ sciences because of ‘deception’. Atoms and particles do not attempt to hide, conceal or deceive their observers. Human beings can and do. Strategic deception should always be a consideration in intelligence analysis, especially when dealing with a foe aware of interest in their activity and have counterintelligence capability.

Hall of Mirrors

When applying a detached, academic analysis to a problem, how can one account for strategic deception by the enemy? The annals of intelligence are filled with tales of strategic deception. Prior to the bombing of Pearl Harbor, Japanese diplomats continued to conduct negotiations, not allowing U.S. analysts to narrow the field of Japanese intentions. Operation Mincemeat saw the body of an RAF officer carrying fake plans released into Spanish waters by submarine and successfully duped German intelligence into believing the allies would invade Greece rather than Sicily. Operation Bodyguard supported the pre-conceived German view that D-Day would come at Calais, not Normandy.

Some deception debates still continue today. Col. Oleg Penkovsky, a GRU military intelligence officer, was the highest ranking source the West had in Moscow during the Cold War. The intelligence he provided on Soviet nuclear capabilities is claimed to have directly influenced President Kennedy’s actions during the Cuban Missile Crisis. However, he was compromised, arrested, put on show trial and apparently executed. Questions remain. Penkovsky’s job did not give him access to the material he provided, so how did he get it? At what point was he compromised and how? Was he executed as a spy or was it a Soviet deception operation from the beginning?

Strangely, after his execution the CIA went to unprecedented lengths to get Penkovsky’s story out, granting full access to the authors of the generously-named The Spy Who Saved the World. Was it to exploit the psychological effect of such a high-ranking source against the USSR? Or was it to cover up the fact they had been comprehensively duped themselves? All this to say that strategic deception is a vital consideration in intelligence analysis and part of James Jesus Angleton’s ‘hall of mirrors’. Is this what they are doing? Or is it what they want me to think they are doing?

While Kentian objective analysis can lead analysts down blind alleys due to strategic deception, Straussian analysis has led in some equally undesirable directions. The values-based judgment Straussian analysis invites can lead one to look at regimes such as the Soviet Union or Saddam Hussein’s Iraq in light of worst-case scenarios because of their (supposed) almost diametrical opposition to the U.S. regime. Those who place primacy on a qualitative analysis of regimes and their ideology and view another regime as possessing an opposing ideology come to view that foe as a major threat where an objective analysis of the same regime considering its past actions, current posture and capability might see the same foe as posing only a minor threat.

From Cuba to Iraq

In 1976, President Ford approved an exercise in competitive analysis, pitting a junior team of CIA analysts applying standard Kentian methods against an external team applying Straussian methods to an analysis of Soviet nuclear capabilities. The episode has come to be known as ‘Team B’. Team B accused CIA of ‘mirror imaging’—assuming the foe holds the same universal principles and goals as the analyst does. Its analysis focused on Soviet intentions, ideas, aspirations and motivations rather than capabilities. It drew conclusions as to intentions based upon an assumption that the Soviet regime not only wanted to defeat the U.S. militarily, but destroy its regime politically, socially and economically. Team B began its analysis from that anchor point and used data to reinforce the assumption as opposed to drawing conclusions from data.

After the fall of the Soviet Union, Team B’s analysis of Soviet capabilities was conclusively proven to be flawed. As an example of how far off path such methods can lead, Team B cited the lack of proof that the Soviet Union had developed an advanced submarine detection system as proof that it did in fact have it. They found it hard to believe the Soviets had not yet done so, so assumed that they had. The continued application of this kind of thinking is exhibited by Defence Secretary Donald Rumsfeld’s 2002 statement on Iraq’s weapons of mass destruction that ‘the absence of evidence is not evidence of absence.’

However, Kentian analysis has also failed at times, most famously in failing in successive National Intelligence Estimates, utilising the entirety of America’s intelligence machinery, to predict the placement of Soviet nuclear missiles on Cuba, leading to the 1962 Cuban Missile Crisis. In 1964, Sherman Kent authored his own well-written post-mortem of how it came to pass, since becoming a classic of intelligence literature. He writes that one of the hazards of the profession of intelligence analysis is filling in gaps in fact with informed guesswork. These gaps are filled with conjecture based upon past actions. That Khrushchev would place missiles on Cuba exhibited a different posture from any of his past actions. Khrushchev ‘zigged’ when they expected him to ‘zag’. Kent goes so far as to say CIA got the estimates wrong because Khrushchev had made such a poor decision that it caught them by surprise.

Returning to the Straussian argument about the Kentian susceptibility to strategic deception, Khrushchev clearly intended to surprise the United States with nuclear missile placements on Cuba and to use them as a bargaining chip against President Kennedy. By suddenly changing posture in a way out of character for Soviet leadership, the Soviet Union was able to fool CIA analysts. Applying Straussian thinking to the situation may have led analysts to assume the USSR would attempt to place missiles on Cuba at some point because of its strategic intent to destroy the United States and look for data to support the assumption. U.S. imagery intelligence did discover missile placements on Cuba, but only because DCI John McCone had suspicions about Khrushchev’s intentions and told them to keep watch over Cuba despite Sherman Kent’s estimates. If estimators began with the assumption Khrushchev would try, they may have been discovered earlier.

This argument about which method of analysis is best suited to producing strategic intelligence estimates has gone on since the beginning of the U.S. Intelligence Community. As episodes from WWII through the Cold War to the Iraq War show, there is no sign of it being decided any time soon. The purpose of intelligence is to speak the truth to power, but the job of analysts is to determine what the ‘truth’ will be before it happens. Whether thinking like Sherman Kent or Leo Strauss, attempting to peer into the future while surrounded by the hall of mirrors is a task anyone is bound to fail at from time to time.