In February 1989, the Soviet Union withdrew from Afghanistan after ten years of brutal counterinsurgency warfare. International observers and Afghan rebels expected the swift collapse of the newly orphaned Afghan communist regime in Kabul, as did the regime itself. Looking to take advantage of the government’s weakness, the Afghan mujahideen temporarily put aside their differences and planned an ambitious assault on Jalalabad, the country's second most important Pashtun stronghold. In March, the rebels rallied a force of some 10,000 soldiers and marched on the city.

They failed spectacularly. Bolstered by Soviet arms, training, and advisers, the Afghan army and its allies easily outgunned, outmaneuvered, and outgeneraled the oncoming force. Although the mujahideen had mastered guerilla tactics and defensive maneuvers on their home turf, they proved incompetent at carrying out a conventional offensive; poor leadership and factional disputes undermined their fighting power. Three years later, the communist regime finally did collapse -- not because it was overrun by the superior fighting abilities of the Afghan rebels but because Russia stopped funding its security forces.

There is no shortage of experts who warn of the impending demise of the current Afghan government after the withdrawal of most NATO military forces at the end of this year. In 2012, the International Crisis Group, for example, warned of a possible “state collapse.” The military historian Tom Ricks wrote that Kabul’s fall was “all too likely.” Yet the case of 1989 suggests the opposite outcome: the Afghan government is likely to survive the withdrawal of international troops, just as the communist regime did, and it stands a good chance of surviving so long as international donors keep the Afghan army in the field. Although Afghans should expect some decrease in international aid, donors -- many fearful that state collapse in Afghanistan could trigger instability in Pakistan -- are unlikely to end their military assistance anytime soon.

The more appropriate historical analogy to Kabul today is not Kabul in 1992; it is Saigon in 1963, when the Vietnamese military overthrew the country’s civilian leadership. In the next five years and beyond, then, the greatest danger to democratic governance in Afghanistan is not the Taliban but the Afghan army itself. Since 2001, the net effect of the state-building mission in Afghanistan has been the emergence of a strong Afghan army and a weak Afghan state, creating an imbalance that should worry anyone with a cursory knowledge of history or political theory. Time and time again, the same recipe has led to countless military coups: highly professional military officers lose patience for corrupt civilian leaders and, educated to believe they are guardians of the state, seize control of the government to save it, often with public support. Coups may sound old-fashioned, but such was the pattern in Egypt and Mali as recently as 2012.

The same could very well happen in Afghanistan, leaving South Asian stability in the hands of a military autocrat -- an outcome that yielded poor results in Pakistan under General Pervez Musharraf, General Zia ul-Haq, and others before them. U.S. and international policymakers should therefore take a number of steps to “coup-proof” the Afghan government. For one, Washington should continue to support an ethnically diverse Afghan officer corps. It could also explicitly condition reconstruction assistance on the maintenance of civilian authority, much as the Kerry-Lugar bill did for U.S. assistance to Pakistan in 2009. The United States need not threaten to withdraw all aid in the case of a military coup, but could threaten to reduce financial assistance, halt the transfer of certain advanced weapons systems, or stop cooperating with Afghan army leaders.

In addition to relying on sticks, however, the United States and its allies must also be able to offer carrots. To that end, Washington needs to sustain a small military presence in Afghanistan, continue investing in relationships with high-ranking Afghan military and civilian officials, provide a steady stream of aid to the army and to the government, and commit to the long and slow process of socializing Afghanistan in the norms of responsible democratic behavior.


The Taliban are an oddity in the history of insurgencies. The group began as a government before deteriorating into an insurgency. Both in and out of power, the Taliban never developed a capacity to capably govern. During their five-year tenure in Kabul, the Taliban proved singularly inept at governing, with Afghanistan ranking in the bottom three states on all of the World Bank’s governance indicators. (By such measures, Afghanistan was governed about as effectively as Somalia, which lacked a government entirely.) As a result of its incompetence and brutality, the Taliban was unable to expand its support among the Afghan population beyond a core group of Pashtun Islamist supporters, instead relying on rule by fear, bribery, and force. And it espoused an unpopular and foreign ideology derived from Deobandi Islam, rather than the traditional Hanafi Islam familiar to most Afghans.

After the 9/11 attacks, the Taliban chose to side with al Qaeda, thereby ensuring that Afghanistan would be visited by yet more war. Once the United States and its allies removed it from power, the Taliban refused to reconcile or participate in the electoral process. Instead, it chose to launch a violent insurgency. But unlike successful insurgents in the past, the Taliban never built up an effective political arm or provided meaningful governance in the areas it controlled.

The Afghan army, meanwhile, has grown increasingly capable and effective. It assumed lead responsibility for the country’s security last summer, meaning that it will have 18 months of experience leading the fight before the withdrawal of most international troops at the end of 2014. Late last year, the U.S. Department of Defense said that the Afghan security forces were conducting almost all operations independently and judged that the Taliban was “tactically overmatched” by Afghan forces, which now number more than 350,000 (including police and local defense forces).

Observers often invoke the legend of the Afghan warrior to explain why international forces have failed to defeat insurgents there. Yet there is no reason why the same legend cannot apply to the regular Afghan army as well. The Northern Alliance, for example, never surrendered to the Taliban in five years of fierce conventional fighting during the 1990s. After a decade of international investment, including some $60 billion in U.S. security assistance and training from the world’s most professional military forces, the Afghan army is far better trained and equipped than the Northern Alliance ever was. Most important, it is likely to have continued air support from the U.S. Air Force.

Given such realities, there is virtually no chance that the Taliban will overwhelm and defeat the Afghan army in battle, march on Kabul, or spur a popular revolution in its favor. The Taliban’s longevity says more about the Afghan government’s weakness and Pakistani complicity than about Taliban strength. Its only chance at outright victory is if the United States and its allies follow in Russia’s footsteps and stop funding the Afghan army. If that were to happen, the army would likely fracture along ethnic lines, giving the Taliban the opportunity to ally with Pashtun units and seize control of the south and east as springboards for national conquest. But U.S. President Barack Obama has already made clear his commitment to providing long-term military and financial support to Afghanistan. So long as he does, the Taliban may continue its sporadic, small-scale attacks, but it won’t win.


The biggest long-term danger to democracy in Afghanistan is not the Taliban but the Afghan army itself, which could overturn the country’s weak civilian government. There would be nothing particularly unusual about a military coup in Afghanistan; military rule, after all, is one of the oldest forms of human political organization. But why do coups happen when they do? Scholars have tended to distinguish between the events that trigger coups, such as the personal ambitions of military leaders or particular political crises, and the underlying structural features that make coups possible in the first place. The first set of variables are unpredictable and, presumably, universal: Afghan officers are as likely to be as ambitious as those in any country, with perhaps more cause to be disgruntled than most. And, of course, Afghanistan has no shortage of crises that officers could use to their advantage. The second set of variables distinguishes countries that are especially vulnerable to a coup. This is where the picture in Afghanistan gets especially worrying: almost every model of coup risk suggests that Afghanistan faces the very real danger of a military takeover.

For example, the British political scientist Samuel Finer suggested that the ideal window for a military to launch a coup is in wartime or during a national emergency -- precisely when the civilian government in most dependent on the military and the military is most popular among the people. Militaries almost always justify their coups by accusing their erstwhile civilian masters of a level of incompetence, corruption, or criminality that endangers the state. That sounds a lot like Afghanistan; the country is in the middle of a protracted war in which the civilian government is entirely dependent on the Afghan army, and during which the government has faced widespread accusations of corruption and incompetence. Last year, Afghanistan ranked 175th on Transparency International’s Corruption Perceptions Index -- the third-worst in the world. It is easy to imagine Afghan corps commanders and the general staff refusing to take orders from Kabul or undermining its control over defense finances.

Another prominent model of coup risk focuses on the size of military establishments. A 1983 study of African coups in the mid-twentieth century by the political scientists Thomas Johnson, Robert Slater, and Pat McGowan concluded that, in states where the military was relatively large and “central because of its role in repression and because of its claim on state revenue,” generals were more likely to become active in civilian politics. That, again, sounds rather similar to Afghanistan. Kabul’s defense budget dwarfs all others. The government collects only about $2.3 billion in domestic revenue, whereas Washington committed $5.1 billion to the Afghan Security Force Fund in 2013. The defense budget is so large, in fact, that it constitutes nearly a quarter of Afghanistan’s $22 billion GDP. With such a large defense establishment, the Afghan army has more resources at its command than the rest of the government put together.

Such an imbalance in resources will naturally give the military greater political influence in shaping Afghanistan’s war strategy, defense posture, and eventually its foreign policy -- allowing a coup by evolution. If civilian leaders resist such influence, they risk precipitating a crisis in which the army overtly seizes control. And it could. There are 185,000 troops under the command of Afghanistan’s Ministry of Defense and over 350,000 Afghan security personnel in total, more than enough to easily oust the country’s parliament, occupy the palace, seize the most important governors’ mansions, and impose martial law in Kabul and Kandahar.

According to the same study, the history of coups in Africa also suggests that rapid population growth in a capital city makes a nation more susceptible to a military coup. Consider, then, Kabul’s population, which has skyrocketed from under one million in 2001 to over 3.3 million in 2013. Inundated with an endless flood of refugees, the city is regularly wracked by protests and civil unrest. A similar period of rioting during Afghanistan's last experiment with democracy in the 1960s led to a widely welcomed coup in 1973.

Other models of coup risk, put forward by the political theorist Rosemary O’Kane and others, emphasize economic and social factors, since a military’s accusations of civilian incompetence often turn on the government’s mismanagement of the economy. Poverty, inequality, recession, high inflation, and volatile commodity markets can leave populations frustrated with their governments. And even growth can have downsides when it is too rapid, disrupting traditional social networks. In Afghanistan, economic and demographic factors are not favorable. The country remains one of the poorest in the world -- even though it is one of the fastest-growing. Afghanistan had the 14th lowest GDP per capita in the world in 2013, but the 15th highest average GDP growth rate from 2001 to 2014. Such figures suggest that Afghans are at once poor and experiencing major disruptions in their way of life. Given these circumstances, a nationalist military with a traditionalist appeal could easily win over Afghans eager for structure and familiarity.

Structural factors also make the Afghan state susceptible to a military takeover. The political scientist Edward Luttwak has argued that three preconditions must be met for a coup to be successful. First, political participation must be limited to a small fraction of the population. (Voting does not count as meaningful participation; countries that hold elections, such as Pakistan, still experience coups because their civil societies are weak relative to their military establishments.) Second, foreign influence must be limited. And third, the state must be centralized. Afghanistan has a weak civil society; it will have less foreign influence after 2014; and its government is famously one of the most highly centralized in the world because the republican constitution gave the president many of the same powers that belonged to the former monarchy. A general who ousted the Afghan president, then, could quickly assume dictatorial powers.

Perhaps the risk factor on which there is the widest agreement is the public’s perceptions of the government’s illegitimacy alongside admiration for the army. A study conducted by the scholars Aaron Belkin and Evan Shofer in 2003 reviewed 21 purported causes of coups and concluded that the three most important variables were the strength of civil society, the level of the regime’s legitimacy, and the past history of coups in the country. This is the most worrying model, because the Afghan state has grown weaker just as its army has grown stronger.

According to the World Bank, the post-Taliban government has improved Afghanistan’s rankings on free speech (not hard considering the Taliban’s repression), government effectiveness, and regulatory quality. Even in these categories, however, Kabul remained at or near the bottom tenth percentile of countries surveyed. In the remaining categories, including political stability, control of corruption, and the rule of law, the democratic government is indistinguishable from its Taliban predecessor, stuck near the bottom of world rankings. Only 47 percent of Afghans say that they trust the Afghan parliament to do its job, 45 percent trust government ministers, and 43 percent trust the court system. A vast 77 percent of them identified corruption as a major nationwide problem.

Meanwhile, in a survey conducted by the Asia Foundation last year, 93 percent of Afghans said the Afghan army was “honest and fair,” 91 percent believed it had helped improve security in Afghanistan, and 88 percent said they had confidence in the institution. Remarkably, these numbers have been largely consistent since at least 2007. Afghans identified the army as the second most honest institution, after the public health-care system. Given Afghan history and its political culture -- coups overthrew the government in 1973 and 1978 -- it would be easy for frustrated Afghan army officers to convince the public that it needed to take over for the civilian government.

By nearly every measure, then, Afghanistan is at high risk of a military coup in coming years. Some scholars and policymakers, however, might argue that a military government would actually be more effective at waging a counterinsurgency campaign against the Taliban and cooperating with U.S. counterterrorism goals. A strong government capable of enforcing law and order could deprive the Taliban of the only real issue on which it credibly claims superiority over a democratic government. By this logic, the international community should quietly wait for the looming coup and continue business as usual after it happens.

That would be a monumental mistake, not least because the Afghans prefer democracy to the alternatives. They continue to express support for political participation, equal rights under law, the freedoms of speech and press, and elections. In the very long run, their support for democracy is a net gain for the United States, Pakistan, and the world. Established democracies tend to be more stable and more economically productive than authoritarian regimes, produce fewer refugees, experience fewer famines, and are less likely to harbor terrorists. And studies suggest that democracies impose fewer costs on their neighbors and on the international community.

Even putting aside the long-term benefits of democracy, which may take decades to realize, accepting a military government in Kabul would be shortsighted in the near term. During the Cold War, Washington often sided with autocratic regimes in the name of defending democracy. The policy succeeded in stemming Soviet influence, but it undermined U.S. credibility, bred anti-Western sentiment in the developing world, and may have indirectly contributed to the rise of nationalist, sectarian, and extremist political movements among populations disillusioned with the rhetoric of democracy.

Moreover, relying on strongmen makes U.S. foreign policy dependent on a single personality. Such rulers generally lose credibility as time wears on and inevitably leave office against their will. Washington would similarly lose out in Afghanistan if, after helping the Afghan army defeat the Taliban, it lost the support of the Afghan people. If Washington wants long-term influence in Afghanistan -- something vital to stabilizing Pakistan -- it must prioritize its relationship with the Afghan people, not just a few favored military officers.


Because Washington has an interest in deterring a coup in Afghanistan, it should push a series of “coup-proofing” policies in the months and years to come. Many states that were otherwise at high risk for a coup, such as Colombia and Iraq, have successfully held their military establishments at bay through a canny mix of patronage, international alliances, ethnic balancing, and support for redundant security organizations. Although some of these tactics are problematic for democracy, if they are pursued in combination with other measures to bolster the government’s legitimacy, they can ultimately help sustain democracy in Afghanistan.

One of the most effective coup-proofing strategies is the exploitation of ethnic and sectarian loyalties. Civilian regimes must either depend on a single ethnic community capable of defending the regime against the rest of the nation or else incorporate a mix of the most powerful groups -- one that could overwhelm other communities while maintaining a balance of power within the armed forces. The first approach is often repressive, undemocratic, and unreliable, for it ensures the permanent dominance of one ethnic group over all others. The Syrian government’s attempt to rely exclusively on the Alawite community demonstrates the danger of that strategy. The Afghan government might, in principle, depend on a military staffed entirely by ethnic Pashtuns, who may constitute 40 to 45 percent of the population. But Tajiks, Uzbeks, and others who fought against the Pashtun-dominated Taliban are unlikely to accept a Pashtun-dominated army.

Fortunately, the international community has already set the Afghan army on the second course, ensuring that its ranks are ethnically mixed. Tajiks are overrepresented, constituting 39 percent of the officer corps, well above the official target of 25 percent. But their overrepresentation comes at the expense of the smallest groups, including Uzbeks and Hazara, not Pashtuns. Pashtuns made up 42 percent of the officer corps and 43 percent of the enlisted ranks, just below the official target of 44 percent.

The army’s balance between Tajiks and Pashtuns is one of the most important guarantors of its viability as a national fighting force under democratic authority. Pashtun officers keen to launch a coup would be unlikely to get support from Tajik officers, and vice versa. Ethnic diversity in the senior ranks may even help deter coups because it would be difficult for a Pashtun cabal, for example, to coordinate without senior Tajik officers getting wind of it. The ethnic balance in the army also carries a risk -- if the army were to fragment, it would split along ethnic lines. But if the military is to remain subordinate to civilian authority, its ethnic diversity is an essential bulwark against a military takeover. Afghan, NATO, and U.S. policymakers should keep a close watch on appointments and promotions to senior positions to monitor the army’s ethnic balance.

A second coup-proofing strategy is to support competing security services. If the regular army plots a coup, it can be deterred by a national guard, a police force, or an intelligence service. The Afghans already have the beginnings of a redundant security system; its national police acts much like a gendarmerie under the command of the Interior Ministry, although it still lacks the funding, training, and equipment required to buffer the army. The Afghan government has also raised local defense forces to defend against Taliban attacks, but they are still too small and poorly equipped to check the regular army. Perhaps the most effective security service outside of the army is the National Directorate for Security, Afghanistan’s intelligence agency. Although it is not a fighting force, its large network of surveillance and intelligence operatives is probably the best resource available to the civilian government to detect disaffection or mutinous plotting within the army. U.S. and Afghan officials should therefore ensure that the NDS devotes at least some of its resources to watching the Afghan army in addition to the Taliban.

A third means of deterring coups involves allowing international influence on a state’s security affairs. The presence of foreign forces from countries devoted to civilian control over the military introduces more uncertainty into the political and security environment (Will foreign forces defend the regime?) and makes it harder for potential coup plotters to know what might happen if they attempt a coup. Foreign funding of a state’s military also gives international donors the power of conditionality over the army.

The presence of external actors and foreign funding is no guarantee of stability, of course, as the coup in South Vietnam in 1963 attests, but it introduces an extra obstacle that coup plotters must overcome. In this respect, Afghanistan is particularly well positioned. The 2012 strategic partnership agreement and the proposed Bilateral Security Agreement have already laid the groundwork for an enduring international military presence in Afghanistan. (The BSA is almost certain to be signed later this year, once a new president takes power, as every candidate has pledged to accept it.) The Afghan army might still dare a coup if it believed that Washington was unwilling to stop it or that U.S. policymakers felt there was no other choice for denying safe haven to al Qaeda. But a military presence surely gives Washington more leverage than otherwise.

Governments must also avoid or end crises that the military could use as a pretext for intervention. In Afghanistan, that means resolving the war against the Taliban in a way that is satisfactory to the country’s Pashtuns while addressing the legitimate concerns of the various non-Pashtun minorities, such as the Tajiks. This is one of the trickiest issues for the Afghan government and its international partners; the most likely coup scenarios would involve Pashtun officers moving against a government that they believed was oppressing Pashtuns in the course of the war against the Taliban -- or Tajik officers against a government that they believed had compromised too much in its eagerness to secure peace. The longer the war lasts after 2014, the greater the chances that tribal grievances outweigh loyalty to the regime. A negotiated solution to the war is a plain necessity, but securing support for such a solution from all of Afghanistan’s major stakeholders will be critical to the survival of the democratic regime.

Finally, the most important measure a state can take to minimize its coup risk is to bolster its own legitimacy. Militaries tend to launch or support coups because they genuinely feel that the survival of their nation or their way of life is at stake. The good news here is that, although Afghans have little confidence in their government and mistrust most of its officials, they continue to support the idea of democracy. Despite some skeptics’ warnings that democracy is unlikely to take root in Afghan soil, the Afghans have proved remarkably patient, and many still hope that democratic norms will eventually take hold.

Since the Afghan government’s dwindling legitimacy is due more to its poor performance than its democratic ideology, the international community can help repair some of the damage. Although much of the burden of cracking down on corruption and nepotism falls on Afghan officials, the international community can equip them with much-needed resources and support. Afghanistan has a weak government and strong army in large part because that reflects the pattern of funding and assistance from the international community since 2001. Most of the international aid to Afghanistan has taken the form of security and economic assistance -- training the army and police and completing high-profile, big-ticket reconstruction projects. A comparatively small amount of money has been spent on the civilian government, and most of that has been spent on elections. Yet Afghanistan needs money for the unglamorous and tiresome work of training bureaucrats, which is the work that might make the largest difference in Kabul’s governance and its legitimacy. As Washington and its allies draw down their military presence, they should increase their support for governance reform and capacity building. Improved government performance will remove the most common pretext for military intervention in politics.

Too often, the United States has ignored the danger of military coups and military governments in allied and partner states. In nearly every case, Washington has found that although military governments can make things easier in the short term, democratic governments allow for a stronger and deeper bilateral relationship. The list of countries taken over by military establishments allied with or trained, equipped, or funded by the United States is embarrassingly long. It includes at least the Dominican Republic, Egypt, Greece, Haiti, Nicaragua, Pakistan, South Korea, South Vietnam, Thailand, and Turkey. Due in large part to the massive investment of U.S. time, money, and resources in the Afghan military since 2001, and to Washington’s relative neglect of the civilian government, Afghanistan is facing a very real risk of becoming the latest entry on that list. There is still time to forestall that outcome. But if it happens, no policymaker -- American, Afghan, or European -- should be surprised.

You are reading a free article.

Subscribe to Foreign Affairs to get unlimited access.

  • Paywall-free reading of new articles and a century of archives
  • Unlock access to iOS/Android apps to save editions for offline reading
  • Six issues a year in print, online, and audio editions
Subscribe Now
  • PAUL D. MILLER is a political scientist at the RAND Corporation. From 2007 to 2009, he was Director for Afghanistan and Pakistan on the U.S. National Security Council.
  • More By Paul D. Miller