Cold War Architecture · Part 5 of 11 · Patreon members only

Founding the Watchers: CIA, NSA, NRO and the 1947 Architecture

Reading time: ~15 minutes

The Problem the 1947 Act Was Solving

Pearl Harbor was the proximate cause. The attack on December 7, 1941 had not been a failure of information — the United States had significant warning indicators available in the signals and diplomatic record. It was a failure of integration. The Army and the Navy maintained separate intelligence operations, shared selectively, competed institutionally, and produced no common picture. The FBI under J. Edgar Hoover ran domestic intelligence as a personal fiefdom. The State Department maintained its own analytical operation. No one was responsible for synthesizing the whole.

The wartime solution was the Office of Strategic Services, the OSS — created by executive order in June 1942 under William “Wild Bill” Donovan. The OSS ran human intelligence collection, covert operations, and analytical work in a single organization modeled partly on Britain’s MI6. It was effective, controversial, and politically radioactive. Harry Truman, who regarded Donovan’s empire with deep suspicion, disbanded it in September 1945, ten days after Japan’s surrender. He scattered its functions among State, War, and Navy.

Within two years, that decision looked badly wrong. The Soviet threat was consolidating in Eastern Europe. The wartime alliance was disintegrating. There was no mechanism for producing a coherent national intelligence picture, and no organization to collect the clandestine intelligence that picture would require. Congress and the newly created National Security Council agreed: the United States needed a permanent, professional peacetime intelligence capability. The National Security Act of 1947 was the legislative vehicle.

The Central Intelligence Agency: Created in Public, Designed for Secrecy

The CIA was created by Section 102 of the National Security Act. It was the most visible intelligence creation of 1947 — the only one that required a public law — and it was designed from the beginning to be something Congress and the public could see while the more sensitive collection apparatus operated elsewhere.

The Act gave the CIA five functions: advising the National Security Council on intelligence matters, making recommendations for coordinating the intelligence activities of other departments, correlating and evaluating intelligence and disseminating it within the government, performing such “services of common concern” as the NSC determined, and performing “such other functions and duties related to intelligence affecting the national security as the National Security Council may from time to time direct.” That last clause — the “such other functions” language — was deliberately vague. It became the legal basis for covert action.

The first Director of Central Intelligence was Rear Admiral Roscoe Hillenkoetter, appointed in May 1947. He presided over an agency that was still assembling itself from OSS remnants when the Soviet Union blockaded Berlin in June 1948. The CIA’s early years were defined less by intelligence collection than by covert action — funding anti-communist political parties in Italy and France, supporting resistance networks in Eastern Europe, running what became a significant paramilitary and propaganda infrastructure.

The collection mission — what intelligence professionals call HUMINT, human intelligence — took longer to build. Recruiting and running agent networks in denied areas requires years of relationship development, language capability, and infrastructure. The CIA spent much of the late 1940s and 1950s building the clandestine service that would become its defining capability. The Directorate of Operations, which runs human intelligence collection and covert action, is the part of the CIA that the public imagination associates with the agency. It is also the part that has produced the most significant collection successes and the most significant operational failures.

The National Security Agency: Created in Secret, Never Debated

The NSA does not appear in the National Security Act of 1947. It does not appear in any public law at all. It was created by a classified presidential memorandum — National Security Council Intelligence Directive (NSCID) No. 9 — signed by President Truman on October 24, 1952. For the first year of its existence, most of official Washington did not know it existed.

The NSA was built on the foundation of the Armed Forces Security Agency (AFSA), itself only four years old, which had inherited the signals intelligence functions of the wartime Army and Navy cryptologic operations. The AFSA had failed at its core mission — coordinating military signals intelligence collection — largely because the individual service branches refused to subordinate their own collection activities to joint oversight. The Brownell Report, commissioned by Truman in 1952, found the situation “unsatisfactory” and recommended a new, stronger organization under direct presidential authority. Truman accepted the recommendation and signed the founding directive.

The NSA’s mission is signals intelligence — the collection and processing of information from electronic signals. This covers communications intelligence (COMINT, the interception of voice and data communications), electronic intelligence (ELINT, the collection of radar emissions and electronic warfare signals), and foreign instrumentation signals intelligence (FISINT, the monitoring of foreign weapons systems during testing). The NSA is also responsible for information assurance — protecting US government communications from collection by adversaries. In this sense it has a direct conflict of interest built into its charter: the same expertise that enables collection also enables protection.

The scale of NSA’s operations was not publicly known until the 1970s, when the Church Committee investigations revealed the scope of domestic collection programs including COINTELPRO and Operation SHAMROCK — a decades-long program in which major US telegraph companies gave the NSA copies of virtually all international telegraph traffic entering or leaving the United States. The Snowden disclosures in 2013 updated that picture for the digital era, revealing collection programs that had scaled with the internet in ways SHAMROCK’s architects could not have imagined.

The National Reconnaissance Office: Created Twice as Secret

If the NSA’s founding was obscure, the NRO’s was extraordinary. The National Reconnaissance Office was created on August 25, 1960, by a classified agreement between the Secretary of Defense and the Director of Central Intelligence. It did not exist as far as the public record was concerned. Its budget was hidden inside the defense appropriations. Its workforce — drawn from the Air Force, the CIA, and civilian contractors — did not acknowledge their employer. Its name was itself classified: NRO employees could not say where they worked.

The NRO was created to solve a specific and urgent problem. The U-2 reconnaissance aircraft had been spectacularly successful since 1956, providing overhead imagery of Soviet military installations at a quality no other source could match. It had also been spectacularly fragile. On May 1, 1960, Soviet air defenses shot down Francis Gary Powers over Sverdlovsk. The political fallout collapsed the Paris Summit and made it clear that overflights of Soviet territory by manned aircraft were no longer sustainable. The satellite programs that would replace the U-2 — CORONA, GAMBIT, and their successors — needed a permanent home. The NRO was built to be that home.

The NRO’s existence remained officially classified until September 18, 1992, when the Department of Defense announced it publicly for the first time — 32 years after its founding. Even after that announcement, specific program names, collection capabilities, and operational details remained classified. The NRO today operates the nation’s reconnaissance satellite fleet, managing systems that collect imagery intelligence (IMINT), signals intelligence (SIGINT), and measurement and signature intelligence (MASINT) from orbit. The ground-based infrastructure that commands those satellites and receives their data — the Aerospace Data Facilities and relay ground stations described in Part 7 of this series — is the NRO’s operational nervous system.

How the Three Agencies Fit Together

The CIA, NSA, and NRO are not redundant — they collect against different targets by different means, and their products feed a common analytical process. Understanding the division of labor clarifies why all three exist and why removing any one of them would leave a gap the others cannot fill.

The CIA collects from people. A well-placed human source inside a foreign government, military organization, or terrorist network can provide context, intent, and decision-making insight that no technical collection system can replicate. Signals intelligence tells you what someone said. Human intelligence can tell you what they meant, what they plan to do next, and what they believe their adversary intends. The CIA’s Directorate of Operations recruits, develops, and manages those sources. It also runs covert action — influencing events abroad through means short of open military force — which is a distinct mission from intelligence collection but uses the same clandestine infrastructure.

The NSA collects from signals. Every electronic transmission — voice call, text message, email, radar emission, telemetry from a weapons system during testing — is potentially collectible. At the scale the NSA operates, signals collection provides volume that human intelligence cannot match. It also provides technical detail — the specific parameters of a foreign radar system, the encryption algorithms protecting a communications network, the RF signature of a new missile — that human sources rarely have access to. NSA’s collection is largely automated; its analytic challenge is processing and making sense of the volume of material the collection systems generate.

The NRO collects from orbit. Satellites provide persistent, all-weather coverage of fixed and mobile targets that no ground-based or airborne system can replicate at scale. A KH-11 optical satellite passing over a military installation produces imagery detail measured in centimeters. An ELINT satellite in geosynchronous orbit can intercept radar and communications emissions from an entire continent simultaneously. The NRO does not analyze what it collects — it operates the platforms and delivers the raw product. The National Geospatial-Intelligence Agency (NGA) processes imagery; the NSA processes signals collected by NRO satellites.

The Intelligence Community That Grew Around Them

The CIA, NSA, and NRO are the three largest and most capable elements of what became the US Intelligence Community (IC) — but they are not the only elements. The IC today comprises eighteen organizations, a structure that reflects seven decades of accretion as new threats, new collection methods, and new policy requirements created demand for new organizations.

The Defense Intelligence Agency (DIA) was created in 1961 to consolidate military intelligence production across the Army, Navy, Air Force, and Marine Corps — the same fragmentation problem the NSA was created to address in signals intelligence. The National Geospatial-Intelligence Agency (NGA), originally the National Imagery and Mapping Agency, was established in 1996 to process and analyze the imagery collected by NRO satellites and other platforms. The Office of the Director of National Intelligence (ODNI), created by the Intelligence Reform and Terrorism Prevention Act of 2004 following the 9/11 Commission’s findings, was designed to provide the coordinating function the 1947 Act had attempted to give the CIA’s Director but which had never been fully achieved. Each of these organizations, and the others in the IC, will be covered in subsequent articles in this series.

The relationship among the eighteen IC components is governed by the same tension that has always characterized American intelligence: the tension between centralization — which produces coordination and eliminates redundancy — and decentralization — which produces competition, diverse analysis, and institutional resilience. The 1947 Act tried to resolve that tension by creating a central intelligence agency. Seventy-eight years of subsequent history suggest the tension is structural, not solvable.

The 1947 Architecture and the UKUSA Agreement

The National Security Act and the UKUSA Agreement were created within eighteen months of each other and are functionally inseparable. The UKUSA Agreement, signed in March 1946, created the signals intelligence sharing architecture. The National Security Act created the US organizations that would operate that architecture on the American side. Without the CIA, there was no US partner for MI6’s human intelligence liaison. Without the NSA, there was no US partner for GCHQ’s signals intelligence. Without the NRO, the satellite collection platforms described in the UKUSA framework’s implementation — the ECHELON ground stations, the satellite-to-ground relay architecture — had no American operator.

The architecture that results from combining these two legal frameworks — one public, one secret — is the system within which global intelligence collection currently operates. The CIA collects human intelligence and shares it with Five Eyes partners under liaison arrangements that run parallel to the formal UKUSA signals intelligence relationship. The NSA is the US node of the ECHELON network. The NRO operates the collection platforms whose product feeds both the US IC and, through carefully controlled sharing mechanisms, Five Eyes partners who have access to imagery intelligence products.

Understanding the 1947 architecture is not historical background. It is the operating framework within which every significant intelligence event of the last eight decades has occurred — and within which the events of the next decades will occur as well.

What This Means for the Prepared

The 1947 architecture was designed to collect intelligence on foreign governments and military threats. Its practical scope, as the Snowden disclosures made clear, extends considerably beyond that original mandate. Several structural realities are worth understanding clearly.

The architecture is enduring, not episodic. These agencies were not created to solve a temporary problem. They were created to provide permanent national intelligence capability for a country that had decided it could not afford, again, the intelligence failure that produced Pearl Harbor. They have survived administrations of both parties, major intelligence failures including 9/11, significant public controversy including the Church Committee and Snowden, and every geopolitical transition of the last eight decades. The collection infrastructure operates continuously regardless of what happens in Washington.

The legal framework governs domestic collection, not foreign collection. The CIA is prohibited by law from conducting domestic intelligence operations. The NSA is required to obtain legal authority before collecting on US persons. The NRO operates in space, where no domestic/foreign distinction applies. These constraints are real and are enforced — but they apply only to collection within the United States and on US persons. Foreign collection faces no equivalent statutory constraint. The architecture’s reach is bounded domestically by law and externally by capability.

The distinction between collection and analysis matters operationally. The CIA, NSA, and NRO are primarily collection organizations. They gather information. The analytical products that reach policymakers are produced elsewhere — by the DIA, the NGA, the CIA’s Directorate of Analysis, and dozens of smaller analytical elements. A collection program that produces significant raw take does not automatically produce significant intelligence unless the analytical capacity exists to process it. This is a structural vulnerability that has been identified repeatedly, including in the 9/11 Commission’s report.

See Also

• The UKUSA Agreement: How Five Eyes Actually Works (Part 4)
• The Five Eyes Satellite Intelligence Network (Part 3)
• America’s Secret Eyes (Part 2)
• Human Intelligence (HUMINT)
• Signals Intelligence (SIGINT)
• Imagery Intelligence (IMINT)
• OPSEC: Don’t Become the Target
• How We Watch: The FFTP Intelligence Collection and Production System

Bibliography & Further Reading

1. National Security Act of 1947, Pub. L. 80-253, 61 Stat. 495 (July 26, 1947). Full text: dni.gov
2. Central Intelligence Agency, “Origins of the Central Intelligence Agency.” CIA History Staff. cia.gov
3. National Security Agency, “NSA/CSS History.” nsa.gov
4. National Reconnaissance Office, “NRO History.” nro.gov
5. Brownell, George A. “A Report Concerning the Communications Intelligence Activities of the Government of the United States” (the Brownell Report), 1952. Declassified. National Security Archive, George Washington University.
6. Church, Frank et al. “Final Report of the Select Committee to Study Governmental Operations with Respect to Intelligence Activities” (the Church Committee Report), 1976. Senate Report 94-755. intelligence.senate.gov
7. Weiner, Tim. Legacy of Ashes: The History of the CIA. Doubleday, 2007.
8. Bamford, James. The Puzzle Palace: A Report on America’s Most Secret Agency. Houghton Mifflin, 1982.
9. Bamford, James. Body of Secrets: Anatomy of the Ultra-Secret National Security Agency. Doubleday, 2001.
10. Day, Dwayne A. “The Development and Improvement of the Corona Satellite.” In Eye in the Sky: The Story of the Corona Spy Satellites, ed. Day, Logsdon, and Latell. Smithsonian Institution Press, 1998.
11. 9/11 Commission. “The 9/11 Commission Report: Final Report of the National Commission on Terrorist Attacks Upon the United States.” Government Printing Office, 2004. 9-11commission.gov
12. Office of the Director of National Intelligence. “Members of the IC.” dni.gov