How espionage really works, beyond MICE and SADRAT

How espionage really works is a question that textbooks answer with tidy diagrams. The SADRAT cycle maps the stages of recruiting a source; the MICE acronym — money, ideology, coercion, ego — sums up why people betray. Both are useful thinking tools. But declassified files, trial records and the memoirs of former officers tell a messier story. Several of the most damaging betrayals of the Cold War did not come from a patient seduction at all: the traitor simply turned up. And many famous failures were not the spies' fault, but that of the people reading their reports.

This is not a manual. Using public sources, it sets out what can be stated with confidence, what we know only from leaks and testimony, and what remains beyond the reach of any outside observer.

Three levels of certainty

What is proven: many of the worst moles walked in

A walk-in is someone who volunteers to a foreign service unprompted — sometimes, quite literally, by walking into an embassy. The most closely studied Cold War files show that several of the most harmful cases were of exactly this kind.

Aldrich Ames (CIA, 1985–1994)

A CIA counterintelligence officer in financial trouble, Ames offered himself to the Soviet embassy in Washington in April 1985. According to the 1994 report by the US Senate Intelligence Committee, he was paid more than $2.5 million and compromised dozens of sources, several of whom were executed in the USSR. He was arrested in February 1994. His case is also a story of internal failure: for years he lived a lifestyle his salary could not explain.

Robert Hanssen (FBI, 1979–2001)

An FBI agent, Hanssen first approached Soviet military intelligence in 1979 and resumed spying in 1985 with a letter to the KGB. He never told his handlers who he was. The Justice Department Inspector General's 2003 review describes how the Bureau's internal security missed, for two decades, warning signs that looked obvious in hindsight. He was arrested in 2001 and died in prison in 2023.

Adolf Tolkachev (USSR, 1977–1985)

A Soviet aviation radar engineer, Tolkachev tried repeatedly to make contact with American diplomats in Moscow before the CIA — wary of a KGB provocation — agreed to run him. The material he handed over was judged of immense technical value. He was betrayed by Edward Lee Howard, a former CIA employee, and executed in 1986. Journalist David E. Hoffman reconstructed the case from the agency's own declassified cables.

Vasili Mitrokhin (KGB, 1992)

A KGB archivist, Mitrokhin spent years copying files by hand. After the Soviet collapse he first approached American diplomats in the Baltics, who did not take him seriously, and then the British. His archive, published with historian Christopher Andrew, remains a key source on Soviet operations — though critics point out that we only see it through his own selection.

The counter-example: the Cambridge Five

Patient recruitment does exist. In the 1930s Soviet intelligence enlisted young British graduates — Philby, Burgess, Maclean, Blunt and Cairncross — on ideological grounds and let them rise for years until they held sensitive posts. It is the textbook case that makes the textbook models make sense. It is not the only pattern, nor, judging by the cases we know, the most common one.

Ideology mattered enormously in the 1930s; in the late Cold War, money, resentment and ego mattered more.

Editorial summary drawing on Andrew and Mitrokhin, The Sword and the Shield

Why MICE falls short

Motives rarely come alone. Ames had debts, but also contempt for his employer; Hanssen took money, but seemed to relish deceiving his colleagues. In 2013 a CIA instructor, Randy Burkett, proposed an alternative framework in the agency's journal Studies in Intelligence: RASCLS, drawn from the psychology of influence — reciprocity, authority, scarcity, commitment, liking and social proof. His point was not that MICE was wrong, but that it explains why someone betrays rather than how a human relationship is built.

The same applies to the SADRAT cycle covered in a separate article: it is a teaching model for understanding stages, not a description of how every case unfolds. Walk-ins, for instance, skip the early phases entirely.

The daily grind: paperwork, patience and suspicion

• Paperwork: former officers describe reports, expense claims and long waits — a world away from the cinema.
• Missed contacts: meetings that never happen and messages that never arrive run through memoirs and trial records.
• Fear of provocation: the Tolkachev case shows a service may turn away a genuine source for fear of a trap.
• Institutional paranoia: James Jesus Angleton's mole hunt inside the CIA in the 1960s wrecked careers and stalled operations without finding the traitor it was looking for.

The most common failure: the leader who won't listen

Many disasters blamed on intelligence were really failures at the receiving end. Three are well documented:

• 1941, Operation Barbarossa: Stalin received numerous warnings of an imminent German attack, including from his own services, and dismissed them as provocations or British disinformation.
• 1973, the Yom Kippur War: Israel's Agranat Commission found that military intelligence clung to a prior 'concept' that Egypt would not attack, despite contrary signals.
• 2003, Iraq: the Robb-Silberman Commission concluded that much of the case on biological weapons rested on a single source, 'Curveball', handled by Germany's BND, who proved to be a fabricator — and that doubts about him never reached policymakers with real force.

False flags: thinking you work for someone else

One approach documented in recent trials is to let a source believe they are dealing with a consultancy, a think tank or a company, when they are really dealing with an intelligence service. In the US, former CIA officer Kevin Mallory was convicted in 2018 after being contacted on LinkedIn by someone claiming to represent a Chinese think tank. In 2020 Singaporean Jun Wei Yeo pleaded guilty to setting up a sham consultancy for Chinese intelligence to obtain reports from American professionals. Britain's MI5 has run public awareness campaigns such as 'Think Before You Link' about approaches through professional networks.

What leaks show — and what remains unknown

Leaks such as Edward Snowden's or Vault 7 have mostly revealed the technical side of intelligence — interception and hacking tools — rather than the recruitment of people. Human intelligence leaves less paper and is guarded more jealously, because exposing a method can cost a life.

That is why sweeping claims such as 'most spies are volunteers' deserve caution. What we know comes from cases that were discovered and prosecuted; the ones that worked, or were never detected, appear in no statistics. It is survivorship bias: we see the failures of whichever side lost the source.

Conclusion

Real espionage looks less like a flawless game of chess and more like a human enterprise full of chance, opportunism and error. Models such as MICE or SADRAT help us think, but the archives teach something else: the decisive factor is usually a person — the traitor who volunteers, the analyst who doubts, the leader who will not listen.