The Paradoxical Nature of Intelligence Organizational Reforms

Key Insight

Efforts to reform intelligence organizations in the wake of failures often fall into a cyclical pattern, swinging between centralization and decentralization, with each approach introducing new problems. The 'creeping determinism' phenomenon not only makes past events seem inevitable but also drives a zeal to 'correct' perceived past problems, inadvertently creating new issues. For instance, after Pearl Harbor in 1941, widely seen as an organizational failure due to scattered intelligence and inter-service rivalry, the Central Intelligence Agency (CIA) was created in 1947 to centralize intelligence collection and processing.

However, just two decades later, the Bay of Pigs fiasco in 1961 was attributed to 'groupthink' within a small, highly cohesive group, suggesting centralization was now the problem. This led to a counter-philosophy valuing 'constructive rivalry,' exemplified by President Franklin D. Roosevelt's approach of assigning overlapping missions to competing agencies and strong personalities to ensure arguments and options 'boiled up from below.' This philosophy shaped the pre-September 11 intelligence community, with the FBI and CIA designed to operate with a degree of rivalry.

Post-September 11, the Shelby Report criticized the FBI and CIA for their rivalry and lack of information sharing, advocating for recentralization, a 'central national level knowledge-compiling entity,' and removing the FBI from counterterrorism, arguing that law enforcement and intelligence 'think differently.' This led to the formation of the Terrorist Threat Integration Center in 2003, consolidating antiterrorist activities. Yet, this reform ignores potential benefits of the previous system: the FBI's specific, 'evidence-supported narratives' (like the Phoenix memo and Zacarias Moussaoui's case) offered valuable insights precisely because they diverged from traditional 'big picture' intelligence analysis. Furthermore, rivalry can function as 'marketplace rivalry,' driving agencies to perform better. Every 'improvement' in intelligence systems involves a trade-off; for example, making warning systems more sensitive reduces surprise but increases false alarms, which then decrease overall sensitivity, demonstrating that a perfect intelligence system remains an elusive goal.