Key Takeaways
Build a political endgame before covert victories reshape the battlefield
A limited mission expands. In 1979, President Carter authorized modest, nonlethal aid to Afghan rebels. After the Soviet invasion, Washington escalated the operation, but its initial purpose was to make Soviet occupation costly, not to design a durable Afghan government. The CIA routed aid through Pakistan, whose leaders saw the war as a way to secure their own frontier and influence Kabul. Those goals were never identical to America’s, even when both sides fought the Soviets.
Success changed the question. By the mid-1980s, the program had grown into a massive arms pipeline, while American officials still treated Afghan politics as someone else’s problem. The central danger was not simply that covert action might fail. It could succeed militarily while leaving the political settlement to allies with different priorities.
The book’s central warning resembles a familiar problem in political science: sponsors often cannot control the local partners who carry out their policies. A proxy is not a tool with a switch. It has its own constituency, history, and definition of victory. The CIA’s restraint about building an Afghan state was understandable after Vietnam, but avoiding political responsibility did not make political consequences disappear. A useful test for any covert campaign is therefore not only whether it can weaken an enemy, but who gains power if it works.
When America delegated the Afghan war, Pakistan chose the winners
The pipeline had gatekeepers. Pakistan’s intelligence service, ISI, controlled the distribution of most American and Saudi supplies. CIA station chief Howard Hart accepted this arrangement because Pakistan controlled access to Afghanistan and Washington wanted to avoid running the war directly. Funding rose from about $30 million in fiscal 1981 to roughly $200 million in fiscal 1984, with Saudi Arabia matching American contributions.
Efficiency had a political price. ISI favored disciplined Islamist clients, especially Gulbuddin Hekmatyar, while Ahmed Shah Massoud received little support. The CIA knew that Pakistan was shaping the resistance but largely accepted its choices because Hekmatyar’s forces fought effectively. After the Soviets left, that relationship gave Pakistan strong influence over Afghanistan’s future, even when its preferred outcomes clashed with American interests.
Delegation can look like a way to save money, reduce risk, and use local expertise. Here it also created a principal-agent problem: Washington supplied resources but lacked reliable control over how its partner used them. The book makes this concrete through weapons distribution, not merely diplomatic disagreement. The complication is that ISI was not irrational or simply corrupt. Its choices reflected Pakistan’s fears about India, Pashtun nationalism, and strategic depth. Recognizing that logic explains the alliance; it does not excuse ignoring what the alliance was building.
A proxy jihad’s networks survived the Soviet enemy they were built to fight
The war built more than armies. Saudi charities, Pakistani intelligence, and Afghan factions helped create schools, guesthouses, supply routes, and training camps for the anti-Soviet jihad. Osama bin Laden first appeared within this world as a wealthy fundraiser and construction organizer. He helped build roads and facilities, supported Arab volunteers, and worked with Abdullah Azzam’s Office of Services, a network that assisted foreign volunteers joining the Afghan war.
The mission outlasted its target. After the Soviet withdrawal, some Arab veterans and organizers redirected their attention toward governments they considered illegitimate, including the United States and Saudi Arabia. Bin Laden’s public profile grew after the 1987 battle at Jaji, and al Qaeda took shape around his expanding ambitions. The CIA had noticed the Arab networks, but it had no sustained policy for addressing them.
The important unit of analysis is not just the fighter or the organization, but the infrastructure that lets people, money, ideas, and skills move between conflicts. Networks created for one purpose can be repurposed when that purpose disappears. This is a sharper account than saying the CIA simply created al Qaeda: the book describes several overlapping sponsors and private actors, with distinct motives. Still, the infrastructure mattered. A campaign can defeat its named enemy and leave behind capabilities that future movements inherit.
Back independent local leaders before proxy partners erase them
Massoud was hard to manage. Ahmed Shah Massoud built a resistance rooted in the Panjshir Valley, using local councils, intelligence sources, guerrilla tactics, and occasional negotiation. He survived repeated Soviet offensives and even arranged a truce in 1983 to protect civilians and rebuild his forces. His independence made him effective, but it also made him unpredictable to the CIA and unacceptable to Pakistan’s ISI, which preferred Islamist clients.
The imbalance compounded. Massoud often received little of the official weapons pipeline, while Hekmatyar’s network benefited from Pakistani support. The CIA later paid Massoud directly, but cooperation remained narrow and conditional. In 1996, when Gary Schroen met him again, the proposed opening centered partly on recovering Stinger missiles rather than on the broad political and military support Massoud wanted against the Taliban.
Massoud’s story shows why local legitimacy can be a strategic asset, not a sentimental bonus. His ties to people in the territory gave him information, resilience, and a reason to stay when other leaders operated from exile. Yet the book does not turn him into a flawless democratic hero. His forces committed abuses, and his government failed to unify Afghanistan. The sharper lesson is comparative: an imperfect independent partner may still be less dangerous than a more efficient proxy whose goals and loyalties are controlled by a rival state.
A Soviet withdrawal was not peace, and Washington mistook it for closure
The predictions were wrong. U.S. analysts expected Najibullah’s government to collapse soon after Soviet troops left Afghanistan in February 1989. It survived for three more years, in part through Soviet aid and political maneuvering. Meanwhile, Pakistani intelligence and the CIA backed a failed assault on Jalalabad that killed thousands of young Afghan fighters without taking the city.
The aftermath was foreseeable. Washington had no agreed plan for a broad postwar government and largely deferred to Pakistan. When Najibullah’s regime finally fell in 1992, rival factions fought over Kabul. The CIA’s covert authority ended, and the United States rapidly withdrew its attention. The country was left with armed factions, devastated infrastructure, and no effective central settlement. A military exit had been mistaken for the end of the problem.
This is a classic case of policy path dependence: the tools used to win one phase of a conflict shape the options available in the next. Once the Soviet withdrawal became the measure of success, diplomacy and reconstruction seemed like distractions rather than parts of the same strategy. The book also shows that prediction errors were not evenly distributed. Analysts were confident about Najibullah’s imminent collapse despite competing evidence and warnings. Forecasts should be treated as revisable judgments, especially when they justify abandoning a country to armed competitors.
The Taliban offered order, while foreign patrons strengthened its absolutism
The opening appeal was real. In 1994, Kandahar was plagued by checkpoints, extortion, kidnapping, and rape. The Taliban gained support by promising security and by disarming local warlords. Their rise was not merely an outside invention. But Pakistan’s military and intelligence services supplied money, equipment, advice, and access to weapons, while Saudi officials and charities offered support and religious connections.
Washington misread the bargain. Pakistan’s leaders hoped a Taliban government would open trade routes and strengthen their influence. Some U.S. officials also saw the movement as a possible source of stability. Unocal promoted a pipeline across Afghanistan, giving commercial hopes added weight in Washington. After the Taliban captured Kabul in 1996, they imposed severe restrictions, especially on women, while hosting bin Laden. American engagement continued, but it did not produce a clear strategy to constrain either the Taliban or their foreign allies.
Order is politically potent after years of chaos, which helps explain why the Taliban could win support without being moderate or broadly representative. The danger lies in treating immediate security as proof of future restraint. Commercial projects can create incentives for peace, but they can also distort attention when a business deal becomes the most tangible reason to engage. The book’s account invites a useful distinction: a movement may provide public goods in one moment while building an authoritarian system that creates deeper insecurity.
Treat terrorist networks as systems, not isolated financiers or criminals
The threat changed shape. The 1993 World Trade Center bombing, the 1995 airline plots, and attacks linked to Afghan-trained militants pointed toward a mobile Sunni extremist network. Yet U.S. agencies initially searched for a state sponsor or treated suspects as separate criminals. The FBI and CIA held different evidence, and legal rules often limited what they could share. The CIA’s Counterterrorist Center was also small and focused on several threats at once.
Recognition came in fragments. By 1996 the CIA had created a small unit devoted to bin Laden, and it gathered reports from Sudan, North Africa, and elsewhere. Analysts increasingly understood that bin Laden financed and connected militants across borders. But the picture remained incomplete, while investigations, intelligence collection, and military planning operated in separate channels. The problem was not a total absence of warnings. It was difficulty turning scattered clues into shared priorities and action.
The network model helps explain why geography-based agencies struggled with an adversary that crossed borders, borrowed charitable and commercial channels, and used both formal organizations and loose personal ties. It also clarifies why a financier should not be dismissed as peripheral: money can connect people who never share a command structure. At the same time, warnings are not the same as knowledge. Many reports were fragmentary or self-serving, and agencies needed evidence that could survive scrutiny. Better coordination matters, but it cannot eliminate uncertainty.
You cannot uproot al Qaeda while shielding the networks that sheltered it
The strategy stayed divided. The CIA developed plans to capture bin Laden, while the White House debated missile strikes, legal authority, and civilian casualties. A proposed raid on Tarnak Farm was rejected amid doubts about its risks. In 1999, a possible strike on a hunting camp was abandoned because members of the United Arab Emirates royal family might be present. Meanwhile, Pakistan’s intelligence service continued supporting the Taliban, and the United States kept hoping that diplomatic pressure would change their behavior.
The final warnings accumulated. The CIA described an expanding al Qaeda network and issued increasingly urgent alerts. Yet Washington would not commit to a broader policy against the Taliban or its Pakistani sponsors. On September 9, 2001, al Qaeda assassins killed Massoud. Two days later, attacks on the United States exposed how closely the Afghan sanctuary and America’s security had become linked.
A government can be tactically active while strategically indecisive. Here, covert plans, missile options, sanctions, diplomacy, and warnings coexisted without a shared answer to the basic question of who sheltered the enemy. Each instrument had real risks, especially for civilians and regional stability. But preserving every alliance also had costs. The case illustrates a hard decision principle: risk cannot be eliminated by postponing choice. Delay distributes risk too, often toward people with the least power to shape policy.
Analysis
Ghost Wars is a history of accumulated choices, not a single blunder. Its narrative runs from the 1979 Soviet invasion through the eve of September 11, but the deeper subject is how institutions convert tactical success into political exposure. The CIA and its partners repeatedly pursued narrow objectives: bleed Soviet forces, contain India, protect a monarchy, recover missiles, or arrest one fugitive. Each objective had local logic. Together they produced a regional system in which armed networks could outlive the conflict that funded them.
The book’s strongest insight is that neglect can be active. Washington did not simply forget Afghanistan. It repeatedly chose to let Pakistan and Saudi Arabia define the practical limits of policy, while retaining a public language of self-determination and counterterrorism. This produced a gap between declared goals and the incentives embedded in money, weapons, diplomatic recognition, and intelligence liaison. The Taliban’s rise was neither wholly indigenous nor merely engineered abroad. It drew strength from genuine popular exhaustion with warlord predation, then gained military reach from external patrons.
The account also resists a simple morality play. Massoud was unusually capable and independent, but his government was implicated in violence and his alliance was not a ready-made liberal democracy. CIA officers faced legal limits, uncertain intelligence, and real civilian risks. Those constraints explain caution without making every delay wise. The most consequential failure was not one missed missile shot. It was the inability to connect intelligence, diplomacy, covert action, and regional strategy around a common definition of the threat.
That distinction matters when reading the book as a causal history of September 11. It does not claim that American aid directly created the attackers or that one alternative policy would certainly have prevented the attacks. It shows how Afghanistan became a sanctuary through repeated decisions by Afghan, Pakistani, Saudi, and American actors, and how the United States recognized danger faster than it could agree on what confronting it required. The title’s ghostly quality is institutional: an enemy can remain visible in reports while remaining politically unreal until catastrophe makes abstraction impossible.
Review Summary
Reviewers widely praise Ghost Wars as a masterful, exhaustively researched account of CIA involvement in Afghanistan from 1979 to 9/11, with most awarding it four or five stars. Readers highlight Coll's balanced, non-partisan approach and vivid characterizations of key figures like Ahmed Shah Massoud and Osama bin Laden. Many describe it as essential reading for understanding modern geopolitics, though some note its density and length make it challenging. A few critics view it as an establishment narrative that glosses over deeper causes of terrorism.
People Also Read
Glossary
Operating Directive
CIA station prioritiesAn annual list of intelligence priorities assigned to a CIA station. It guided what officers were expected and resourced to collect. When Afghanistan was removed from the Islamabad station’s directive after the Soviet Union collapsed, the station chief had less formal authority and funding to recruit agents for Afghan political or military intelligence.
mujahedin
Afghan resistance fightersThe term refers to Muslim fighters who joined the Afghan resistance against the Soviet occupation and the communist government in Kabul. The mujahedin were not one unified army. They operated through competing parties, commanders, regional networks, and foreign sponsors, and their rivalries shaped Afghanistan after the Soviet withdrawal.
ISI
Pakistan's intelligence servicePakistan’s Inter-Services Intelligence Directorate, usually called ISI, is the military intelligence service central to the book’s account. During the anti-Soviet war it controlled much of the distribution of CIA and Saudi weapons and money to Afghan factions. It pursued Pakistan’s own goals in Afghanistan and Kashmir, sometimes alongside and sometimes against American priorities.
Stinger
Portable heat-seeking missileThe Stinger is a shoulder-fired, heat-seeking antiaircraft missile supplied by the CIA to Afghan rebels during the Soviet war. It helped the mujahedin threaten Soviet helicopters and aircraft. After the war, the CIA feared that unaccounted-for missiles could reach hostile governments or terrorists, and it funded a costly program to buy them back.
NSDD-166
Expanded Afghan covert policyNational Security Decision Directive 166, signed by President Reagan in 1985, expanded U.S. support for the Afghan guerrillas. It authorized broader intelligence, weapons, training, and operational support than the earlier policy of harassing Soviet forces. The directive made a more ambitious covert war possible while drawing more American agencies into its management.
al Qaeda
Bin Laden's militant networkAl Qaeda is the organization built around Osama bin Laden and his allies, including Ayman al-Zawahiri. It developed from transnational networks of Arab volunteers, fundraisers, veterans, and Islamist organizers connected to the Afghan jihad. The book portrays it as a core leadership group surrounded by wider circles of militant allies, charities, and recruiters.
Counterterrorist Center
CIA counterterrorism hubThe CIA’s Counterterrorist Center, created in 1986, combined operations officers, analysts, and technical specialists to address terrorist threats that crossed regional boundaries. It initially focused heavily on secular groups and later expanded its attention to Islamist networks. Its small bin Laden unit became a central but resource-constrained part of the campaign against al Qaeda.
rendition
Cross-border suspect transferRendition is the transfer of a terrorism suspect from one country to another, often without a formal extradition process. U.S. administrations used it to bring suspects to American courts or to allied countries for interrogation and prosecution. The practice occupied a legally sensitive space between intelligence operations, law enforcement, and foreign sovereignty.
blowback
Unintended return effectIn intelligence usage, blowback describes the return of covert actions or planted information in ways that harm or confuse the country that initiated them. In the book, the KGB spread a false claim that Afghan leader Hafizullah Amin was a CIA agent. The rumor later fed Soviet suspicions that the claim might be true.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.