In February 2025, Israeli Prime Minister Benjamin Netanyahu presented President Donald Trump with an unusual gift: a gold-plated pager. The object commemorated Israel’s extraordinary September 2024 operation against Hezbollah, in which thousands of pagers distributed among members of the organization exploded almost simultaneously across Lebanon. A second wave of explosions struck handheld radios the following day.
As a trophy of intelligence proficiency, the pager strike is difficult to surpass. The operation required a remarkable combination of intelligence penetration, deception, engineering, supply-chain manipulation, operational security, patience, and precise timing. An adversary worried about Israeli penetration of cellular communications had turned to supposedly safer, low-technology devices. Israel apparently penetrated that solution as well. Devices acquired precisely because they were thought secure became weapons against their users.
The “golden pager” presented by Israeli Prime Minister Benjamin Netanyahu to President Donald Trump. Photo: Israel Government Press Office.
One need not minimize the accomplishment to question its significance. Quite the opposite. The operation is analytically useful precisely because its operational virtuosity is so clearly evident. It demonstrated an extraordinary ability to penetrate an adversary’s organization and convert that penetration into coordinated physical effects. Presenting the golden pager to Trump was signaling the extraordinary proficiency of Mossad, and by extension, the power of Israel.
But what, exactly, did that proficiency prove? Could Israeli intelligence reach deeply enough into Hezbollah’s communications supply chain to place destructive devices in the hands of its personnel and activate them at a chosen moment? Clearly yes. But harder questions follow. How much did the operation durably diminish Hezbollah’s military capability? How rapidly did the organization adapt? Did the attack alter recruitment, leadership behavior, communications practices, or political support? Most important, did it make Israel durably safer?
These questions concern something different from technical accomplishment: strategic value. The distinction is easily lost because operational success is so much easier to see. The pagers exploded. The targets were struck. The ingenuity of the operation was understood around the world within hours. Strategic effects arrive on a different schedule. They emerge through adaptation, succession, retaliation, political reaction, changed incentives, and events that may unfold for years. Some may never be confidently attributable to the original operation.
Elite national intelligence services like Mossad have acquired a mystique based upon their ability to accomplish extraordinarily difficult things. Much of that proficiency is real. The mistake comes when demonstrated mastery of clandestine operations is enlarged into a presumption of national strategic benefit. Finding a target is not the same thing as defining a problem. Destroying a target is not the same thing as eliminating a threat. Executing a mission brilliantly is not the same thing as pursuing a sound strategy. A gambit is not a game. The golden pager tells us that Israeli intelligence can hit a target. It does not tell us whether Israel is winning the game. That distinction, between tactical success and strategic efficacy, is the subject of this article.
The Cult of Proficiency
Mossad is hardly unique. National intelligence services occupy a peculiar place in the modern political imagination. The CIA, MI6, Mossad, French intelligence, and their adversary counterparts are not regarded merely as government bureaucracies. They are mythologized as elite orders of professionals operating beyond the competence, and often beyond the comprehension, of ordinary institutions.
The mythology contains enough truth to make it durable. Intelligence work can demand extraordinary skills. Recruiting a source inside a hostile government, penetrating an encrypted network, identifying a concealed weapons facility, or conducting a clandestine operation in hostile territory can require patience, technical mastery, ingenuity, courage, and meticulous organization.
But proficiency possesses a peculiar gravitational force. Once demonstrated in one domain, it attracts assumptions about competence in others. The intelligence professional knows things we do not know. He sees dangers invisible to us. He understands how the world really works. Much of his best evidence cannot be shown because revealing it would compromise sources and methods. The outsider therefore encounters an unusual proposition: the very secrecy that prevents independent evaluation becomes part of the reason for deference.
This produces what might be called an intelligence proficiency cult. The term does not imply that the underlying proficiency is imaginary. It describes the migration of admiration for difficult execution into assumptions about broader understanding and judgment.
Operational proficiency is unusually easy to recognize. A source was recruited. A facility was destroyed. A hostage was rescued. A leader was killed. A network was penetrated. The accomplishment has an objective, a beginning, an end, an identifiable actor, and frequently a measurable result.
Strategic judgment has none of these advantages. Whether an assassination weakened an organization may remain uncertain for years. Whether sabotage delayed a weapons program enough to matter requires a counterfactual judgment. Whether overthrowing a government improved the geopolitical position of the country that overthrew it may be disputed for generations.
The spectacular covert operation therefore enjoys an enormous presentational advantage over the strategic ledger. Former intelligence officials appear as authoritative commentators on diplomacy, war, terrorism, political movements, foreign cultures, and the intentions of entire nations. Their experience may provide unusual insight. But expertise in acquiring secret information or conducting clandestine operations does not automatically confer superior judgment about the political systems in which those operations occur.
The inferential chain of intelligence capability is seductive. They can penetrate the adversary. Therefore, they understand the adversary. Thus, they can predict the adversary. Finally, they know what should be done about the adversary. Each proposition requires evidence that the preceding proposition does not supply.
Popular culture enormously amplifies these inferences. Intelligence agencies possess something most government institutions could only envy: an immense fictional auxiliary. For generations, novels, films, and television have supplied intelligence professionals who know more, see more, endure more, and accomplish more than ordinary people.
James Bond supplies the classic template. The adversary is identifiable, the danger enormous, and the mission comprehensible. Bond penetrates the sinister organization, discovers the plan, overcomes the obstacles, and destroys the threat. The narrative ends because the mission supplies its own strategic resolution. The target was the problem, and eliminating the target solved it.
Jason Bourne offers an even purer distillation. Bourne loses much of his memory and with it his biography, institutional context, loyalties, and identity. Yet the essential quality survives: extraordinary operational proficiency. Strip almost everything else from the intelligence operative and the mythology tells us what remains at the core: proficiency.
There is an instructive irony in the enormous contribution of spy fiction to the intelligence mystique. John le Carré, who did as much as anyone to make the secret world culturally compelling, reached a remarkably deflationary judgment about its actual historical importance. “It wasn’t the spies who won the cold war,” he observed in 1993. “I don’t believe that in the end the spies mattered very much at all.” Nearly a quarter-century later, his judgment was even more categorical: “Spies did not win the Cold War. They made absolutely no difference in the long run.”
Le Carré was hardly an outsider mocking a profession he did not understand. He had worked for both MI5 and MI6 before becoming the preeminent novelist of Cold War espionage. His retrospective judgment captures the distinction between operational drama and strategic consequence. Espionage produced extraordinary stories of penetration, betrayal, deception, courage, and technical ingenuity. It also produced an enormous cultural mythology. But the Soviet Union ultimately collapsed for reasons having little to do with the clandestine contest that had so captivated both sides.
Intelligence agencies benefit from this dubious cultural inheritance. When Mossad penetrates an adversary’s communications network or the CIA finds a fugitive after years of searching, the event arrives in a public imagination already trained to recognize what it means: the professionals demonstrated their prowess and the mission succeeded.
Once an intelligence service accumulates enough celebrated successes, its reputation can approach critical mythological mass. New accomplishments confirm institutional character: Of course Mossad could do that. Failures receive different treatment. They become exceptional episodes requiring particular explanations: a collection failure, an analytic failure, an individual error, a coordination problem, an unlucky contingency. For elite intelligence organizations, success generalizes; failure particularizes.
The result is the category error at the center of the proficiency cult: demonstrated mastery of clandestine operations becomes evidence for a much broader presumption of competence. The intelligence service has shown that it can accomplish extraordinarily difficult things. It has not thereby established that it understands the larger strategic problem, can predict its evolution, or knows what should be done about it.
Gambits Are Not Games
The fundamental error of the intelligence proficiency cult is treating operational accomplishments and strategic outcomes as though they were different degrees of the same thing. They are not. A target is not a problem. A mission is not a strategy. Proficiency is not wisdom. A gambit is not a game.
Consider the difference between darts and chess. In darts the objective is straightforward: hit the target. A bull’s-eye is intrinsically successful. A chess gambit cannot be evaluated that way. Capturing a piece may be executed perfectly and still be disastrous because its value depends upon the result of the game.
Clandestine operations are frequently presented as dart bull’s-eyes when they should be evaluated as chess moves. A scientist is assassinated. A terrorist leader is killed. A nuclear facility is sabotaged. A communications network is penetrated. A government is destabilized. Each operation can be scored as an operational success. But none of those scores tells us what result has been produced on the larger geopolitical chessboard.
The problem begins with target substitution. Complex strategic problems consist of political systems, institutions, populations, technologies, grievances, incentives, cultures, alliances, and adaptive adversaries. Intelligence organizations must make such problems actionable. Nuclear proliferation becomes a scientist, centrifuge facility, procurement network, or shipment. Terrorism becomes a leader, financier, communications node, or training camp. An insurgency becomes an infrastructure of identifiable people and organizations.
This transformation is operationally necessary. Intelligence agencies cannot sabotage “nuclear proliferation.” They must act against something concrete. The mistake occurs when the actionable representation of the problem becomes confused with the problem itself. Killing the scientist does not destroy his knowledge. Killing the commander does not necessarily destroy his organization. Destroying the facility does not necessarily eliminate the program it supported. Overthrowing a government does not determine the political system that replaces it.
This is operational myopia: evaluating an action within a system boundary too narrow to capture the consequences by which its ultimate value should be judged. Within that boundary, the result may be impeccable. The correct individual was identified. The weapon functioned. The target was destroyed. The mission was accomplished. Outside the boundary lie succession, regeneration, adaptation, retaliation, political reaction, escalation, opportunity costs, and altered adversary incentives. These are not peripheral consequences. They are often where strategy is determined.
Leadership decapitation illustrates the problem. Killing a leader may eliminate years of experience and disrupt an organization. But his successor may be younger, more capable, more aggressive, or harder to locate. The organization may decentralize and become harder to penetrate. The dead leader may acquire greater symbolic value as a martyr. Or the successor may be incompetent and the organization may fragment. Decapitation triggers succession, but the attacker does not choose the successor.
A more sophisticated defense invokes cumulative effects. No individual strike need be decisive. Repeated operations may deplete leadership, consume security resources, disrupt communications, and gradually weaken the adversary. That is a serious argument, but it creates an empirical obligation. If cumulative degradation is the theory of victory, the consequential result, not the latest destroyed target, is the appropriate measure.
The proficiency cult encourages the opposite accounting. Operations arrive one at a time. Each target is important. Each successful strike is another accomplishment. Each replacement becomes another valuable target. Repeated operational success can thereby become insulated from strategic falsification: if the strike destroys its designated target, the strike succeeds; if the adversary remains capable of resistance, that persistence establishes the need for another strike.
The logic resembles bloodletting as an historical medical treatment. The treatment demonstrably does something, but the patient remains ill. Instead of challenging the causal theory, persistence of the illness becomes evidence that another treatment is necessary. Operational success combined with strategic insufficiency can produce the same inversion: every successful operation validates the method, while every failure to achieve the strategic objective validates the need for another operation.
This does not mean clandestine operations are strategically futile. Some produce substantial strategic benefits, and some problems really can be solved by finding the right target and acting against it. The analytical requirement is simply that the connection be demonstrated rather than presumed. Intelligence organizations are unusually capable of demonstrating the first half of the proposition: they found the person, penetrated the network, sabotaged the facility, or destroyed the weapon. Strategic validation requires the second half: What happened because they did? The proficiency cult stops at the bull’s-eye. Strategic analysis must examine the chessboard.
From Entebbe to the Golden Pager
There are circumstances in which the dartboard model works reasonably well. Operation Entebbe is an unusually clean example. On June 27, 1976, an Air France flight from Tel Aviv to Paris was hijacked after leaving Athens and eventually diverted to Entebbe Airport in Uganda. The hijackers separated Israeli and Jewish passengers from most of the others and held them while demanding the release of prisoners. On July 4, Israeli commandos flew more than 2,000 miles to Uganda, stormed the terminal, killed the hijackers, and rescued most of the remaining hostages. The commander of the assault force, Yonatan “Yoni” Netanyahu, was killed.
Israeli commandos preparing for Entebbe raid
Entebbe became a foundational legend of Israeli military and intelligence prowess for good reason. It was an extraordinarily difficult operation requiring intelligence, deception, logistical planning, surprise, audacity, and exceptional execution. More important for our purposes, its operational objective and strategic purpose were closely aligned. The hostages were in immediate danger. Their location was known. The adversaries were identifiable. The desired outcome was concrete: rescue the hostages. Accomplishing the mission would largely accomplish the larger objective.
Entebbe therefore provides a useful control case. Operational proficiency can produce decisive strategic value when the problem itself is sufficiently bounded. The danger lies in generalizing the lesson. Entebbe supplies an almost irresistible model of national-security action: identify the threat, penetrate its defenses, strike with precision, and resolve the problem. At Entebbe, the target and the problem substantially coincided.
But Iran is not an airport terminal. Hezbollah is not a group of hijackers guarding hostages in a room. Hamas is not a command post. Nuclear proliferation is not a centrifuge facility. Palestinian politics is not a target set. Lebanon is not a pager communications network.
These adversary challenges are complex adaptive systems containing institutions, populations, technologies, political constituencies, ideologies, grievances, external patrons, accumulated knowledge, succession mechanisms, and capacities for learning. An intelligence operation can perturb such a system; it does not necessarily terminate it.
This is where the difference between Entebbe and the golden pager becomes important. The September 2024 pager attack was an astonishing intelligence feat. Israeli intelligence had apparently penetrated the supply chain through which Hezbollah acquired communications equipment, transforming devices intended to protect the organization from Israeli electronic surveillance into weapons against their users.
But the pager network was not Israel’s strategic problem. Hezbollah’s ability to threaten Israel did not reside inside several thousand pagers. It resided in a much larger political-military system of fighters, commanders, weapons, infrastructure, political institutions, constituencies, accumulated expertise, and external support. The attack could damage that system. The strategic question is how much, for how long, and with what subsequent effects.
Nearly half a century separates Entebbe from Netanyahu’s presentation of the gold-plated pager. Both belong to the same tradition of Israeli operational audacity, but they should not carry the same analytical meaning. At Entebbe, hitting the target largely solved the problem. With the Hezbollah pager attack, Mossad demonstrated that Israel could hit the target. The rest of the strategic ledger remains to be written.
The Ledger Nobody Keeps
If intelligence agencies were ordinary organizations, their reputations would presumably rest on some reasonably complete accounting of performance. Successful operations would count in their favor. Failed operations would count against them. Long-term outcomes would be compared with objectives. Mistaken operating theories would eventually be revised. Unfortunately, intelligence services operate in an environment unusually resistant to such accounting.
Intelligence agency successes and failures have markedly different properties. Successes are often discrete, visible, vivid, and attributable. A hostage is rescued. A fugitive is found. A facility is sabotaged. A leader is killed. Failures are different. They are diffuse, temporally extended, causally complicated, frequently classified, and almost always contestable. Success has an agent: Mossad did it. Failure has a passive voice: warnings were missed.
Intelligence successes can become legendary feats. Failure must be decomposed into causation. Was it a collection failure? An analytic failure? Did policymakers ignore a warning? Did an individual make the wrong judgment? Was the failure systemic? Could anyone reasonably have predicted the outcome? Each question may be legitimate. Intelligence operates against adversaries who conceal, deceive, and adapt. No serious evaluation can demand omniscience. But the cumulative effect of these qualifications is remarkable: intelligence success concentrates praise on the intelligence agency; failure has causes that diffuse and deflect responsibility.
A successful operation demonstrates the potency of the institution. A failure becomes an episode requiring explanation. Success arrives as a simple story. Failure arrives as a complicated discussion. The pager operation can be explained in a headline. Understanding the institutional failures preceding 9/11 requires hundreds of pages of commission findings, timelines, disputed warnings, and arguments about what different agencies and officials knew. The more complicated the failure, the easier it becomes for institutional responsibility to dissolve into explanation.
The agency defenses can form a kind of orderly retreat. An adverse event did not happen. If it happened, perhaps it was not authorized. If authorized, perhaps particular officers exceeded what was intended. If the conduct was institutional, perhaps the failure was systemic. If the system failed, responsibility becomes diffuse. And if the strategic result remains unfavorable, perhaps classified successes prevented something worse. Any one of these defenses can be valid. The problem lies in their aggregate effect. Positive evidence concentrates as institutional credit while adverse evidence can be progressively decomposed until responsibility becomes difficult to locate.
Secrecy makes the imbalance greater. Intelligence agencies necessarily know things the public does not, and legitimate sources and methods may require protection for decades. But uncertainty is not evidence of success. Intelligence services may possess an unknown inventory of classified triumphs. They may also possess an unknown inventory of classified failures. If the contents of the secret ledger are unknowable, they cannot logically be entered only on the credit side.
Counterfactual arguments require the same symmetry. Perhaps an assassination prevented a devastating attack, sabotage prevented a war, or repeated strikes weakened an adversary that otherwise would have become much more dangerous. But perhaps leaving a cautious leader alive would have been preferable to his successor. Perhaps preserving an intelligence penetration would have yielded more value than exploiting it. Perhaps an operation produced retaliation or adaptation worse than the danger it removed. We cannot count the unknowable favorable counterfactual while excluding the unfavorable one.
Claims about cumulative operational effects are more testable. If repeated disruption, attrition, and leadership depletion are supposed to weaken an adversary over time, then the relevant accounting unit is the trajectory.
The CIA Phoenix Program in the Vietnam War provides an important test case. Phoenix and related efforts sought to identify and “neutralize” the Viet Cong Infrastructure through capture, defection, and killing. Whatever disputes surround particular counts and methods, the program operated on a scale far beyond the spectacular leadership assassinations that dominate modern intelligence mythology.
Its tactical effects were real. Official statistics for 1968–1972 reported more than 81,000 members of the Viet Cong Infrastructure neutralized, including roughly 26,000 killed, with the remainder captured or induced to defect. These figures did not represent 26,000 targeted assassinations; many of those counted as killed died in military operations. But by any reasonable measure, Phoenix and the intelligence-driven campaign surrounding it removed large numbers of people from the Viet Cong political infrastructure. Defenders have argued that this damaged the underground and imposed significant organizational costs.
But Phoenix also exposes the limits of body-count logic. If eliminating enough members of a hostile infrastructure cumulatively produces strategic victory, Vietnam supplied an enormous experiment in that proposition. The United States could count the people neutralized, but the result did not enable South Vietnam to survive the eventual withdrawal of American power. A campaign that neutralized tens of thousands of members of the enemy’s political infrastructure did not produce the strategic outcome that its impressive tactical metrics might have implied.
Effective organizations require corrective feedback. If repeated action fails to produce the predicted higher-order result, the theory connecting action to outcome should come under scrutiny. Localized success cannot substitute for general efficacy.
The proper intelligence ledger therefore cannot consist merely of successful missions. It must ask what happened after they succeeded. Did the target organization weaken significantly? Did the weapons program recede? Did the political environment improve? Did the threat diminish? Did the need for further coercive action decline? If so, operational proficiency deserves strategic credit. If not, the theory justifying another similar operation requires reconsideration.
Tactical Focus, Strategic Blindness
The defense against this critique is that intelligence is an inherently uncertain business. Adversaries conceal intentions, plant false information, compartmentalize plans, change direction, and sometimes surprise even themselves. An intelligence service that never failed would not be proficient. It would be omniscient. That is not the standard proposed here. The relevant historic failures are not missed meetings, undiscovered agents, or erroneous satellite interpretations. They involve wars, nuclear confrontation, strategic surprise, penetration of the intelligence service itself, and fundamental misunderstanding of major adversaries. Consider the CIA record.
The Bay of Pigs was not merely an unsuccessful clandestine operation. The CIA helped develop and execute a plan based on profoundly optimistic assumptions about what a small exile invasion could accomplish politically inside Cuba. The April 1961 landing failed catastrophically, humiliating the Kennedy administration and strengthening Fidel Castro’s dependence on the Soviet Union.
Prisoners after failed Bay of Pigs invasion
Its consequences did not end on the beach. Kennedy intensified efforts to remove Castro; Cuba sought greater protection against another invasion. Operation Mongoose followed, and within eighteen months Soviet nuclear missiles were being installed in Cuba. It would be simplistic to say that the Bay of Pigs “caused” the Cuban Missile Crisis. But a clandestine operation intended to solve the Cuba problem became part of a chain of actions and reactions that made the problem vastly more dangerous. That is precisely the larger system boundary within which clandestine operations must be evaluated.
At the other end of the Cold War came a different intelligence challenge. For decades the Soviet Union was the central object of American intelligence collection and analysis. Satellites photographed it. Signals were intercepted. Defectors were questioned. Its military forces, economic performance, and leadership politics were closely studied. Yet the Soviet system entered terminal crisis and disappeared with a speed the American intelligence establishment did not correctly predict. Analysts certainly identified many of its economic, political, and nationalities problems, and political collapse is extraordinarily difficult to forecast. But the contrast remains instructive: an intelligence service may know the location of missile silos with remarkable precision while remaining much less certain about the political viability of the state that owns them. This was operational myopia at continental scale.
The Aldrich Ames case supplies an even less forgiving test because it involved the CIA’s own professional domain. Ames, a CIA officer specializing in Soviet counterintelligence, began spying for the Soviet Union in 1985 and continued for years, compromising American assets, some with fatal consequences, while displaying conspicuous wealth inconsistent with his government salary. The significance is not that a spy penetrated an intelligence service. Every major service faces penetration attempts. It is the contrast between mythology and organizational reality: an institution celebrated for penetrating adversaries was itself penetrated for years by an officer working in the field where it should have been most alert to precisely that danger.
Then came September 11. This was not simply “a CIA failure.” It exposed deficiencies across a larger intelligence, law-enforcement, aviation-security, and policymaking system. Warnings about al-Qaeda existed, as did fragments of relevant information in different places. The system nevertheless failed to convert what it knew into prevention of the most consequential foreign attack on the United States since Pearl Harbor. Collection is only one component of intelligence performance. Information must be recognized as significant, integrated, communicated, and translated into useful warning. More information is not necessarily more understanding.
The Iraq WMD failure was different again. The central problem was not predicting an unknowable future but determining an existing material condition: whether Saddam Hussein’s Iraq possessed the weapons programs attributed to it before the 2003 invasion. The principal judgments were wrong. Disputes about politicization, caveats, dissenting assessments, and policymaker exaggeration matter, but they do not erase the analytical failure. The American intelligence system reached high-confidence conclusions about Iraqi weapons capabilities that subsequent investigation could not substantiate. Those judgments helped provide the justification for an invasion whose enormous human, financial, and geopolitical costs vastly magnified the consequences of the intelligence failure.
Afghanistan supplies another variation. After twenty years inside the country, the United States possessed intelligence officers, military intelligence, surveillance systems, diplomatic reporting, allied relationships, contractors, Afghan institutions, and enormous quantities of data. Yet in 2021 the Afghan government and security forces collapsed with astonishing speed. Withdrawal itself changed incentives, and political systems can cross thresholds rapidly. But after two decades of intimate involvement, the speed of collapse raises an unavoidable question: how well did the United States understand the political and institutional structure it had spent twenty years attempting to sustain?
These cases should not be homogenized. The Bay of Pigs was an operational and strategic failure. The Soviet collapse presented a strategic-warning problem. Ames was a counterintelligence failure. 9/11 was a warning and integration failure across multiple institutions. Iraq WMD was an analytic failure with enormous policy consequences. Afghanistan exposed the limits of long-duration strategic understanding. Their heterogeneity is the point.
If intelligence mythology claimed only that intelligence officers sometimes perform extraordinarily difficult operations with extraordinary skill, none of these episodes would refute it. But the mythology confers something larger: epistemic prestige. Intelligence institutions are presumed to see beneath appearances, understand hidden threats, anticipate adversaries, and provide knowledge unavailable through ordinary means. That larger claim must confront the larger ledger.
An intelligence service may identify which apartment an adversary occupies, which telephone he carries, which vehicle he enters, and when he will arrive at a meeting. It may penetrate a supply chain or sabotage machinery without entering the facility. Yet the same national intelligence system can fail to understand whether an invasion will succeed, whether its principal adversary is approaching collapse, whether terrorists are preparing a catastrophic attack, whether a targeted state possesses the weapons attributed to it, or whether a government supported for twenty years can survive for twenty weeks. That is the paradox concealed by the proficiency cult: extraordinary tactical acuity can coexist with remarkable macroscopic blindness.
Failure-Induced Hypertrophy
The failures described above have another curious feature. They have done remarkably little to arrest the long-term growth of the intelligence institutions responsible for avoiding them. This does not mean intelligence failures are consequence-free. Directors resign. Congressional committees investigate. Commissions issue reports. Procedures are revised. Oversight mechanisms are established. Budgets sometimes decline.
But step back from individual episodes and examine the CIA across its history. The dominant trajectory is difficult to miss. An organization created in 1947 principally to collect, coordinate, evaluate, and disseminate foreign intelligence rapidly acquired covert political-action responsibilities. Those expanded into sabotage, support for coups, paramilitary operations, proxy warfare, and assassination plots. By the twenty-first century, the CIA possessed armed unmanned aircraft and operated globally in finding, targeting, and killing individuals thousands of miles from the United States.
The growth has not been linear. Missions have changed, budgets have risen and fallen, and programs have been terminated. But institutional size and scope have displayed a pronounced tendency to accumulate. A rough chronology illustrates the direction of travel:
The table should not be read as claiming that each failure caused the capability listed beside it. History is not that mechanical. The important feature is the asymmetry: New capabilities are comparatively easy to add. Old capabilities are remarkably difficult to subtract. Failure can even become an engine of growth. An investigation discovers insufficient collection, poor information sharing, inadequate technology, jurisdictional barriers, or too few personnel. Each deficiency generates a plausible remedy: more money, personnel, technology, access, authority, or integration. The diagnosis may be correct. The cumulative institutional effect is nevertheless hypertrophy.
CIA headquarters in Langley, VA
Success provides another route to the same destination. A spectacular operation demonstrates the value of an existing capability, generating prestige, political confidence, and demand for more. Thus apparently opposite feedback signals can produce the same institutional response: Success demonstrates why the agency deserves more power. Failure demonstrates why it needs more power. At some point an intelligence institution can reach critical mythological mass. Capability produces operational options. Operational options increase political reliance. Reliance produces resources and discretion. Resources produce additional capabilities. Success demonstrates their value; failure reveals deficiencies requiring correction, resulting in additional resources and authority.
The resource trajectory confirms the pattern. Publicly disclosed National Intelligence Program appropriations rose from $43.5 billion in fiscal 2007 to more than $70 billion by the mid-2020s, while combined national and military intelligence spending reached roughly $100 billion annually. Earlier CIA budgets remain substantially obscured by classification, but declassified Agency histories describe rapid expansion during its first years.
Headcount alone understates the transformation. A modern intelligence officer supported by satellites, global communications interception, immense databases, cyber tools, artificial intelligence, foreign liaison networks, contractors, and precision weapons represents vastly more capability than an officer in 1950. The revealing historical measure is not simply how many people work for the CIA, but how many things the CIA has acquired the capacity and authority to do.
The long-term institutional-capability curve is strongly upward. The strategic-efficacy curve is much harder to find. After coups, sabotage, proxy wars, assassinations, counterterrorism campaigns, drone strikes, and steadily improving surveillance and targeting, where is the corresponding demonstration of increasing mastery over strategic outcomes? If the benefits are cumulative, the evidence should eventually become cumulative as well.
The history of Mossad suggests that this is not simply an American bureaucratic peculiarity. Israel’s foreign intelligence service likewise evolved from a comparatively modest coordinating and collection organization into one associated with overseas special operations, kidnapping, assassination, sabotage, technological penetration, cyber operations, and close integration with military action. Its mythology and operational repertoire have grown together. Yet despite these extraordinary capabilities and accomplishments, Israel continues to face grave security threats. Operational intelligence success has not translated into strategic resolution.
The accretion of intelligence agency power matters because capability creates options. An intelligence service that can penetrate a network, sabotage a facility, find a commander, or kill him places concrete capabilities before political leaders. Such options possess an inherent appeal compared with diplomacy, deterrence, political accommodation, or institution-building: a clandestine operation can often produce a measurable result quickly. Operational capability therefore begins to shape policy.
The agency remains formally subordinate to political authority. But the relationship becomes less straightforward when it helps identify the threat, controls much of the secret information about it, develops the clandestine options, assesses their feasibility, and possesses the capability to execute them. The deeper danger of intelligence agency hypertrophy is therefore not simply institutional size. It is the progressive alteration of the relationship between intelligence and strategy. An agency created to inform policy can gradually acquire the power to shape it.
Truman’s Frankenstein
The National Security Act of 1947 created a Central Intelligence Agency whose central purposes were recognizable as intelligence functions: coordinating intelligence activities, evaluating national-security information, and disseminating intelligence to policymakers. The new organization was supposed to help the President understand the world. This functional boundary did not last long.
Within the emerging Cold War, intelligence collection became entangled with a different proposition. If an intelligence organization could secretly discover what foreign governments and political movements were doing, perhaps it could also secretly influence what they did. The verbs began to change. Observe, collect, assess, and inform were now augmented by penetrate, manipulate, sabotage, overthrow, capture, and kill.
This was not simply mission growth. It was a transformation in institutional character. The logic was understandable. The Soviet Union and its allies conducted espionage, political warfare, subversion, propaganda, and clandestine operations. American policymakers wanted comparable instruments. Covert action offered something enormously attractive: the ability to influence events abroad while obscuring American responsibility.
By 1948, barely a year after the CIA’s creation, covert political action had been incorporated into the emerging intelligence apparatus. The operational repertoire subsequently expanded to propaganda, political intervention, sabotage, paramilitary operations, coups, proxy forces, and assassination plots. The institution established to inform policy had become capable of secretly implementing it.
Harry Truman eventually regarded this transformation with alarm. In December 1963, the former President wrote that he had become “disturbed by the way CIA has been diverted from its original assignment.” He said he had never intended CIA to become involved in what he called “strange activities” and argued that its operational duties should be terminated or otherwise assigned. Whatever one makes of Truman’s retrospective criticism of programs authorized by his own administration, his warning identifies the institutional problem. The issue is not simply what an intelligence agency knows. It is what an intelligence agency is empowered to do.
Collection and analysis have a natural stopping point: information is delivered to political authority. Covert action collapses that distance. The intelligence organization can identify a problem, develop an operational response, present it to political leadership, and, if authorized, execute it. As operational authority expands, the distinction between adviser and actor begins to blur.
Plausible deniability complicates the relationship further. It was not merely an informal practice but an explicit element of U.S. covert-action policy. NSC 10/2, issued in 1948, defined covert operations as activities planned and executed so that U.S. Government responsibility would not be evident and, if exposed, could be plausibly disclaimed—a principle carried forward in subsequent National Security Council directives. Plausible deniability can serve legitimate strategic purposes by avoiding escalation and protecting agents, foreign partners, technical capabilities, and ongoing operations. But it also weakens the normal connection between authority and responsibility. The operation is important enough for the state to undertake but politically inconvenient for the state to acknowledge.
Clandestine systems therefore make responsibility unusually difficult for outsiders to reconstruct. Intelligence officers can exceed authority, instructions can be ambiguous, operations can develop momentum, and allies can act independently. Retrospective intelligence history contains numerous operations once denied, disputed, or hidden that eventually entered the official record. The lesson is not that every contemporary denial is false. It is that deniability cannot itself establish that the disputed activity did not occur.
The CIA’s transformation eventually crossed an even more consequential boundary. During the post-9/11 War on Terror, surveillance, target development, and lethal action became increasingly integrated. Predator drones initially used for surveillance were armed with Hellfire missiles. The CIA acquired the ability not merely to find targets or provide intelligence to military forces, but to participate directly in killing them. An intelligence agency had acquired an organic aerial strike capability.
The traditional sequence: intelligence service finds → political authority decides → military force strikes could increasingly be compressed within a clandestine architecture integrating collection, targeting, operational planning, and lethal execution. Artificial intelligence, ubiquitous surveillance, cyber penetration, enormous data stores, and automated pattern recognition can compress the cycle further. But greater speed and precision do not resolve the strategic question. Faster identification does not establish that the identified person should be killed. Better target discrimination does not establish that destroying the target advances strategy. AI can optimize the find-fix-finish cycle without determining whether the cycle itself produces the desired political result.
The system can become better at answering how while becoming no better at answering why. This does not make the CIA uniquely sinister. Many of these capabilities arose in response to genuine threats, were authorized by elected governments, and were supported by Congress. The question is institutional. At what point does an intelligence service possessing secret information, covert political capabilities, sabotage expertise, paramilitary forces, technological penetration, and lethal strike capacity cease to be merely an intelligence service and become something substantially more powerful?
Truman’s Frankenstein metaphor is useful not because the creature escaped all control. The more subtle danger is that the creature grew in capability while becoming increasingly indispensable to those charged with controlling it. The intelligence service had not seized political power; political power had become increasingly dependent upon the intelligence service.
Strategic Amnesia
Institutional aggrandizement requires a peculiar kind of memory. Intelligence services must remember operational knowledge extraordinarily well. Sources, networks, technical methods, adversary behavior, surveillance techniques, cryptographic vulnerabilities, and accumulated tradecraft constitute institutional capital. An effective service cannot afford to rediscover these things every generation. Strategic memory operates differently.
The history of covert action is filled with interventions whose long-term consequences proved ambiguous, disappointing, or disastrous. Governments were overthrown only to produce later instability or hostility. Leaders were assassinated and replaced. Infrastructure was sabotaged and rebuilt. Political movements were penetrated but survived. Yet these experiences rarely seem to accumulate into a comparable presumption of restraint.
The contrast with operational learning is striking. If an assassination fails because a detonator malfunctions, the service studies the detonator. If surveillance fails because the target changes communications methods, it develops better surveillance. But if repeated successful assassinations fail to eliminate the organization, the strategic premise behind assassination may receive far less rigorous reconsideration. The lesson becomes: improve the operation.
It does not mean intelligence agencies literally forget their history. They conduct reviews, preserve archives, and study previous operations. The amnesia lies in the failure to convert accumulated strategic experience into constraints on future action. Operational lessons become institutional capability. Strategic disappointments become inert history.
The public environment reinforces the imbalance. Successful operations remain vivid because they can be recounted and retold. Classification is lifted. Participants write memoirs. Journalists uncover technical details. What was once a mysterious explosion becomes an impressive story of surveillance, deception, penetration, courage, and ingenuity. Failure ages differently. Responsibility fragments. Policymakers die. Political circumstances change. Causal arguments become more complicated. Eventually the episode becomes something historians dispute rather than something policymakers fear repeating. The institutional credit survives more readily than the strategic debit.
Political incentives reinforce the same tendency. Intelligence operations provide attractive trophies because they offer visible evidence of decisiveness without requiring strategic resolution. A raid can be announced. A leader can be declared dead. A sabotaged facility can be photographed. A spectacular penetration can be commemorated with a golden pager trophy. Strategic success is much less cooperative. Deterrence produces events that do not happen. Diplomacy may look like inactivity. Political stabilization can take decades. Preventing an adversary from becoming an enemy generates no dramatic footage.
The incentives therefore favor operational action. The politician demonstrates resolve. The intelligence service demonstrates proficiency. The media receive a compelling story. The public receives an identifiable victory. There are no comparable rewards for asking what the operation accomplished five years later. The result is a powerful asymmetry in institutional learning. Operational methods are continuously refined because their performance is visible and their failures are actionable. Strategic theories are harder to test, their consequences emerge slowly, and responsibility for failure is diffuse. The machinery for executing the next operation therefore improves more steadily than the machinery for deciding whether the operation should be undertaken. Tactical proficiency keeps improving. The unresolved question is whether the strategic return improves with it.
Restoring Strategic Discipline
The remedy for the distorting influence of the intelligence proficiency cult is not less proficient intelligence. Nations need clandestine sources, technical collection, counterintelligence, warning, analysis, and sometimes covert capabilities. The problem arises when operational proficiency becomes insulated from strategic efficacy and institutions possessing extraordinary secret powers become increasingly responsible for judging the value of exercising them. Every effective organization requires both accountability and corrective feedback. Secrecy may constrain how accountability operates, but it should not sever the feedback loop between action and result.
The governing principle for reform of intelligence agencies should be straightforward: Operational authority must be governed by strategic purpose, and operational success must be independently tested against strategic efficacy. The principle becomes especially important when clandestine activity involves assassination, sabotage, lethal force, or plausible deniability. Evaluation cannot end when a target dies, a facility is destroyed, or a network is disrupted. It must ask what followed: how the adversary adapted, whether the effect endured, what unintended consequences appeared, and whether the threat that justified the operation materially diminished. These activities need not be prohibited categorically, but they should carry a burden of strategic proof commensurate with their consequences.
That requires a more fundamental reform: mandatory independent review of strategic efficacy. Major clandestine programs and sustained campaigns should be periodically evaluated by institutions that did not propose or execute them. The review should begin not with the number of successful operations but with the causal theory that justified them.
Relevant questions are: What was supposed to happen? What actually happened? Did the predicted strategic effect appear? How did the adversary adapt? Did the threat diminish? Does the accumulated evidence justify program continuation? These are elementary management questions. They provide the corrective feedback essential to any effective organization. Review must also possess a termination function. Repeated operational success accompanied by failure to produce predicted strategic effects must trigger reconsideration of the operating theory, not merely authorization of another similar operation. If cumulative effects are the justification, show us the curve.
There is an important historical precedent for such reform. In 1975–76, the Senate’s Church Committee conducted the most searching investigation of the American intelligence establishment in its history. Its inquiries exposed domestic surveillance, assassination plots, covert interventions, and other activities that had developed behind extraordinary secrecy. The committee concluded that intelligence abuses had flourished in part because normal constitutional checks and balances had not been adequately applied to intelligence activities.
The resulting reforms were substantial. Congress established permanent intelligence oversight committees, imposed stronger reporting requirements on covert action, and enacted the Foreign Intelligence Surveillance Act. Executive orders and Justice Department guidelines imposed additional constraints. Intelligence agencies did not emerge from the 1970s institutionally untouched; an earlier generation had already recognized the dangers created by secret power insufficiently subjected to independent review.
But the Church reforms also illuminate the remaining problem. Oversight can ask whether an operation was legally authorized, properly reported, adequately supervised, and consistent with statutory or constitutional limits. Those are indispensable questions. They are not identical to asking whether the operation worked strategically.
An assassination can be properly authorized and flawlessly executed yet leave the adversary stronger. A sabotage operation can comply with every reporting requirement yet produce only temporary disruption followed by rapid adaptation. A covert campaign can satisfy the machinery of legislative oversight while its underlying causal theory remains wrong.
The next step in intelligence accountability is therefore not to discard the architecture created after Church, but to extend its logic. Independent oversight should test not only whether secret power was exercised lawfully and properly authorized, but whether repeated exercises of that power produced the strategic effects invoked to justify them.
There is also a constitutional dimension. Intelligence organizations combine secret information, specialized expertise, operational discretion, coercive capability, limited public visibility, and privileged access to executive authority. The stronger these characteristics become, the more important independent institutional review becomes.
The ancient Roman Praetorian prefect Sejanus provides an instructive precedent. Under Tiberius, Sejanus accumulated extraordinary influence as the emperor increasingly withdrew from Rome. Control of access and influence over the information reaching Tiberius helped transform a formally subordinate official into an enormously powerful political actor.
Modern intelligence agencies are not Praetorian Guards, and the danger does not require a coup or conspiracy. The mechanism is subtler. An intelligence agency can identify the threat, possess information unavailable to most policymakers, assess its seriousness, develop clandestine responses, advise political leaders about their feasibility, and then execute the option selected.
Formal political supremacy remains intact. Yet informational asymmetry can alter its substance. A president dependent upon an intelligence organization to define the threat, explain the adversary, identify the options, assess their likelihood of success, and execute the selected operation exercises authority within an informational environment substantially constructed by the institution being supervised.
The danger of Praetorian capture is not necessarily that the guard overthrows the emperor. It is that the emperor increasingly sees the world through the guard. The safeguard is not simply stronger executive control, which can deepen dependence upon the secret institutions closest to the executive. It lies in competing sources of information and independent centers of judgment: meaningful legislative oversight, inspectors general with genuine investigative authority, judicial review where appropriate, independent strategic assessment, competing military and diplomatic analysis, and eventual public disclosure when legitimate secrecy requirements expire.
Separation of powers is therefore not merely a constitutional abstraction. It is an information architecture. No institution possessing secret information and clandestine coercive power should dominate the mechanisms by which the necessity and efficacy of that power are judged.
The objective is to restore the distinction between proficiency and authority. An intelligence service may be extraordinarily good at penetrating an adversary without possessing superior judgment about policy toward that adversary. Possession of information unavailable to everyone else makes independent scrutiny more difficult, not less necessary.
No effective organization would allow the unit proposing an undertaking, executing it, measuring its performance, and controlling the underlying data to become the sole judge of whether the undertaking should continue. The stakes are vastly greater when the activities in question are coups, sabotage, assassinations, and war.
Intelligence agencies do not require less competence. They require bounded authority, defined objectives, independent evaluation, accountability, corrective feedback, and abandonment of operating theories that repeatedly fail their strategic tests. Extraordinary proficiency is an argument for stronger strategic discipline, not an exemption from it.
Conclusion
There is no need to deny the extraordinary proficiency of the world’s leading intelligence services. Mossad’s penetration of Hezbollah’s communications supply chain was astonishing. The CIA has recruited agents inside difficult intelligence targets, found fugitives who spent years hiding, penetrated hostile organizations, and developed technical capabilities that would have seemed fantastic to its founders. Intelligence officers sometimes perform difficult and dangerous work with remarkable courage and ingenuity. That is precisely why the distinction between operational and strategic efficacy developed in this article matters.
The mystique associated with the intelligence agencies is not a mistake because the proficiency is imaginary. The mystique is a mistake because operational proficiency is mistaken for expansive competence. Operational accomplishment creates legitimate admiration. Admiration creates prestige. Prestige encourages deference. Deference permits demonstrated skill at clandestine operations to migrate into assumptions about superior understanding, judgment, prediction, and ultimately strategic wisdom.
The historical record gives ample reason to resist that migration. Extraordinary microscopic acuity has coexisted with macroscopic blindness. Spectacular operations have accumulated alongside strategic failures. Intelligence capabilities and authorities have expanded far more visibly than any corresponding improvement in strategic outcomes. This does not establish that clandestine operations are futile. It establishes that operational success cannot validate itself. The proper measure is what happens afterward: whether the adversary weakens, whether the threat diminishes, whether the predicted political effect occurs, and whether the need for further coercive action declines. If cumulative effects are claimed, the cumulative evidence must eventually appear.
The same principle applies institutionally. Secret information, operational capability, and privileged access to political leadership make intelligence agencies unusually powerful instruments of government. Those characteristics increase rather than diminish the need for independent judgment. No organization should become the principal judge of the strategic efficacy of operations that it proposes, executes, and evaluates itself.
Which brings us back to the golden pager. As a trophy of intelligence ingenuity, it deserves its place in a museum. It represents patience, penetration, technical imagination, deception, operational security, and audacity assembled into an astonishing clandestine achievement. But it is not a strategic balance sheet entry. It tells us that Israel could reach Hezbollah. It does not tell us that Hezbollah was defeated. It tells us that a target could be hit. It does not tell us that the target was the problem. That distinction is easy to lose because operational success is spectacular and the strategic ledger is complicated. The danger is that such success becomes an end in itself, detached from the strategic purposes intelligence operations are supposed to serve. The mystique of proficiency is not a substitute for strategic mastery.
