In late 2001,when U.S. forces expelled the Taliban from Afghanistan, the country appeared headed for a breakup. The United States and the rest of the international community feared that Afghanistan's rival ethnic groups would use their regional power bases to pull apart any unitary state, forming in its place independent ministates or aligning with their ethnic brethren across Afghanistan's borders. At the time, such fears seemed credible: NATO troops were still dealing with the fallout from the violent disintegration of the former Yugoslavia in the 1990s.

The Afghans themselves, however, were less concerned about their country dividing. After all, Afghanistan has been a single state for more than 250 years. If the country were going to split, it would have done so in the 1990s, during its protracted civil war. Yet it did not. No Afghan leader of any political stripe or ethnicity endorsed secession at any time during the last century. Nor did any at the start of this one. Although Afghanistan's various ethnic factions disagreed about how the country's new government should be organized and who would wield power within it, they all proclaimed their support for a unitary state.

A decade later, the anxiety of Washington and its allies has reversed itself. If in 2001 the West was afraid that the absence of a strong centralized government in Kabul would prompt Afghanistan's dissolution, by 2011 the West has come to fear that a dysfunctional centralized government could cause this same outcome. Such a turn of events was caused by several factors, perhaps most of all by many Afghans' dissatisfaction with a centralized national administrative structure that cannot cope with the country's regional diversity or with expectations for local self-rule. The government in Kabul has been further undermined by the country's fraudulent 2009 presidential election, the absence of political parties, poor security, and general corruption.

As a result, the fears of 2001 have come to life, as regional and ethnic ties have taken on an ever-larger role in Afghan politics and society. Networks based on ethnicity have become stronger conduits for patronage and protection and have often merged with criminal syndicates that exploit similar ties. After 2005, when the Taliban insurgency began to reenergize in Afghanistan's largely Pashtun southern and eastern regions, part of its appeal was rooted in this local opposition to the dysfunctional government led by Afghan President Hamid Karzai.

Yet despite this reality, the United States and its allies seem not to have considered whether a different configuration of the Afghan state and its leadership -- one created by the Afghans themselves -- might prove more stable and inclusive than the present one. Devolving the political authority to redress the current imbalance between national and local governance is not the first step toward Afghanistan's disintegration but a way to avoid it. Such reforms would solve a dilemma currently standing in the way of peace in Afghanistan: how to create space for dealing with insurgents whose concerns are local rather than national or international. In 2001, Afghans from every region and ethnic group were so eager for peace that they accepted the restoration of a flawed central government. Today, the mood is quite different. Unless the United States and its allies confront -- and resolve -- the problem of political legitimacy before foreign forces leave the country, the West's efforts to create a stable Afghanistan will fail.


Afghanistan's population of 30 million people is divided into seven major ethnic groups -- the Pashtuns, the Tajiks, the Hazaras, the Uzbeks, the Aimaqs, the Turkmens, and the Baluchis -- and many smaller ones. Although the Pashtuns claim to be the national majority, most analysts believe they constitute only a plurality of between 40 and 45 percent. However, each ethnic group does constitute the majority in one or more of Afghanistan's regions: Pashtuns in the south and east,Tajiks in the northeast and west, Hazaras in the center, and Uzbeks in the northwest. On the political level, ethnicity in Afghanistan is more descriptive than operational, as most individuals' primary loyalty is local -- to kin, village, valley, or region. There is little political cohesiveness within large ethnic groups, except when faced with an enemy ethnic group. Meanwhile, crosscutting ties of intermarriage, bilingualism, and political alliance regularly transcend ethnicity. Among non-Pashtuns, shared location is even more important: different ethnic communities in the same towns and villages often display more solidarity with one another than they do with their ethnic compatriots from other parts of the country. As a result, Afghan ethnic groups have never viewed themselves as fixed nationalities with an overriding commonality and history that would require political unity or a nation-state. Instead, ethnicity in Afghanistan is essentially prenationalist, with ethnic groups holding similar economic and political interests but no common ideology or separatist aspirations. Moreover, the multiethnic state had long been the accepted norm in Afghanistan; it was not some anomaly that needed rectification. Ethnic conflicts in Afghanistan historically centered on which group would dominate the state and subordinate others, not over which group would have exclusive control of a territory.

For most of the time between the sixth century BC, when the Persian Empire was founded, and the mid-eighteenth century, Afghanistan was divided among empires based in Central Asia, India, and Iran. These Turkish and Persian ruling dynasties taxed and administered each region's cities, trade routes, and most productive agricultural areas. They won political support by tying themselves to local elites who became junior partners in government. These regimes ignored the poor mountainous and desert regions, since they would not repay the cost of administration, until they caused trouble. (The proud boast that Afghanistan has never been conquered is true only for these remote regions, not its cities and productive lands.) In 1747, a Pashtun dynasty founded by Ahmad Shah Durrani took control of Afghanistan for the first time, but Durrani followed the same pattern as his Turkish and Persian predecessors. Rulers in Kabul appointed governors (often their relatives) to regional cities but let these governors remain largely autonomous.

This pattern changed when Abdur Rahman Khan came to the throne in 1880. Later known as the Iron Amir, Rahman aimed to rule Afghanistan directly, without relying on intermediaries. Before finally subduing the country in 1895, his regime put down more than 40 uprisings and killed more than 100,000 people. He ended the regional political autonomy that had formerly characterized Afghanistan and concentrated all political power in Kabul. The Pashtuns became Afghanistan's politically privileged group, although the Tajiks ran the government's administration. Uzbek, Aimaq, and Turkmen leaders disappeared from public life, even in their home regions; the Hazaras, meanwhile, faced active discrimination. Although all succeeding Afghan regimes emulated Rahman's centralized state, it proved hard to maintain. In the century following Rahman's death in 1901, every one of Afghanistan's rulers either died violently or was driven into exile.


Although ethnic conflict played a role in Afghan politics in the twentieth century, it was never the reason for state collapse. Instead, it was ideology that brought down most regimes. The country was cleaved by an unresolved conflict between the modernizing elites in Kabul and the country's more conservative rural inhabitants. The modernizers assumed that they could change Afghanistan by decree, but they underestimated the military strength and administrative capacity required to accomplish such a task. This conflict first proved fatal in 1929, when the reform-minded Amanullah Khan was overthrown by rural revolts only months after demanding sweeping social and legal changes. Two generations later, in 1978, the same conflict arose when the People's Democratic Party of Afghanistan promulgated radical new social and economic policies that alienated rural Afghans. Only a Soviet invasion the next year and a protracted occupation would keep the PDPA in power.

These conflicts did not pit different ethnic groups against one another: Pashtun kings and Communists opposed rural Pashtun mullahs and tribal khans; a progressive Persian-speaking intelligentsia in Kabul opposed conservative Persian-speaking villagers. In both 1929 and 1978, insurgents allied across ethnic and regional lines, citing the common threat to their traditional ways of life and interpretations of Islam. (Although the current insurgency in Afghanistan reflects similar ideological and cultural fault lines, the Taliban have found it more difficult to transcend their rural Pashtun ethnic base.) In Afghanistan, such ideological conflicts are associated with opposition to established regimes, whereas ethnic and regional conflicts emerge in the vacuum after a state collapse. Despite the decade that has passed since the Taliban were driven from power, Afghanistan, at least politically, still remains in this post-collapse era of uncertainty. Therefore, frustrating the Taliban insurgency is less about confronting the Taliban's ideology, which was never very popular in Afghanistan, than it is about creating a more stable and legitimate Afghan state.

If insurgencies topple established Afghan governments, their stable replacements emerge from the crucible of civil war. Rural rebels may stay united long enough to bring down governments in Kabul but lack enough internal cohesion to create stable regimes of their own. When faced with a common threat or united in a common goal, such groups readily set aside the disputes that ordinarily divide them, only to rediscover them once the goal has been achieved. This dynamic played out most clearly in the 1929 alliance of conservative Tajiks (from north of Kabul) and Pashtun tribes (from eastern Afghanistan) that overthrew Amanullah. These former allies became enemies when Habibullah Kalakani, the leader of the Tajik faction, unilaterally named himself emir of Afghanistan. For the cause of Pashtun solidarity, the very tribes that had just forced Amanullah from the throne now rallied behind him. After nine months, the Pashtun force defeated Habibullah and installed its own commander, Muhammad Nadir Shah, as ruler of Afghanistan. Restoring the old ethnic hierarchy and jettisoning Amanullah's reforms, Nadir and his heirs governed a peaceful Afghanistan for the next 50 years.

On a much longer and larger scale, the mujahideen insurgency against the PDPA that began in 1979 and the civil war that followed that regime's collapse in 1992 matched this pattern. The mujahideen rallied nationwide opposition to the decadelong Soviet invasion of Afghanistan. When the Soviets withdrew in 1989, however, unity among the mujahideen began to fray. Those factions that had aimed to expel the Soviets saw little point in continuing the fight against Najibullah, then the leader of the PDPA. He had already jettisoned the regime's radical communist ideology and now offered arms, money, and local autonomy to those factions willing to leave the resistance. This tactic succeeded well enough until the Soviet Union disintegrated in late 1991, taking with it the outside military and financial support for the PDPA regime. The government collapsed in April 1992. Most former members of the PDPA then joined mujahideen parties based on shared ethnic ties.

The Afghan civil war of the 1990s had no ideological foundation and pitted ethnic- and regionally based factions against one another. With the exception of Abdul Rashid Dostum, the secular Uzbek militia leader, all the major factional leaders (Pashtun, Tajik, Hazara) were Islamists of one variety or another. Although militia leaders mobilized members of their own ethnic groups to fight, they fought not for the cause of an ethnic group but out of self-interest. Conflicts that appeared ethnic on the surface were, in reality, fights over control of political, economic, and military resources. This fact explains the often bewildering shifts in alliances among different militia groups during the civil war and the striking lack of unity within any single ethnic group.

If there was any overarching ethnic theme in the civil war in the 1990s, it was the attempt by non-Pashtuns to break the century-old ethnic hierarchy that had discriminated against them. They demanded a return to an older pattern of regional autonomy, in which local elites played a significant role in governing their own people and had a say in politics at the national level. The Pashtuns bridled at this assertion of power but could do little to stop it. They were too divided and unable to unite behind a single Pashtun leader. In the past, they had always rallied around scions of the royal Durrani house, a good strategic ploy because both rural Pashtuns and non-Pashtuns accepted the legitimacy of such rulers. But in the 1990s, competing Pashtun mujahideen leaders, such as Gulbuddin Hekmatyar, vetoed this option, since each man wanted supreme power for himself. By 1994, in the absence of Pashtun unity, the civil war reached a stalemate: each faction could hold its own territory but was unable to move beyond it.

The Taliban, a movement led by Mullah Omar and other low-level Pashtun clerics from Kandahar, broke this standoff. As Ibn Khaldun, the Arab social historian, first noted seven centuries ago, religious leaders often prove more successful than tribal ones in uniting large groups in Islamic societies. Calling on God's authority, the Taliban circumvented clan rivalries and united people in the name of religion. In theory, this dynamic should have given the Taliban nationwide appeal, but their leadership and membership remained too overwhelmingly Pashtun for that to happen. Although Taliban victories between 1995 and 1998 gave them control over most of Afghanistan's territory, non-Pashtuns resented the regime's heavy-handed tactics and narrow ethnic chauvinism. They waited for an opportunity to turn against the regime. This moment, of course, would arrive in 2001, when the United States went to war against the Taliban in response to the 9/11 attacks. Within ten weeks of the beginning of U.S. military operations, the Taliban were expelled from Afghanistan. In the end, not even the Pashtuns in the Taliban's home base of Kandahar proved willing to fight on their behalf.


The fall of the Taliban left a power vacuum at the national level and restored clout to regional leaders. But none of them opposed the establishment of a new unitary government or sought to break away from it once it had been created. In the months and years that followed the U.S. invasion, the discussions about Afghanistan's future and the architecture of its political system -- the 2001 Bonn meetings, the 2002 national loya jirga (council of elders), and the 2003 constitutional convention -- were marked more by cooperation than by conflict. Equally significant was what did not happen: the non-Pashtun Northern Alliance leaders, who had defeated the Taliban, did not form a government unilaterally; except for some local score settling, the Pashtuns were not penalized for their earlier pro-Taliban sympathies; and no faction took on the role of a spoiler. The agreement to make Karzai the head of state recognized the traditional Pashtun claim to executive power, but the ministries were divided among the different ethnic factions. Although many non-Pashtuns initially opposed Afghanistan's draft constitution because they wanted a less centralized government, the representatives ratified it unanimously because no faction wished to pick a fight. Afghanistan's long tradition of practical politics made these agreements possible; the new government was an arranged marriage, not a love match. Its legitimacy would be judged by its effectiveness. Ironically, the major expected benefit of making Karzai, a Pashtun from Kandahar, president -- winning the support of the southern Pashtuns -- never materialized, because politics among rival clans in the south divided the community between pro- and anti-Karzai factions to the benefit of the Taliban.

The new government had an exceptionally strong executive, modeled on the constitution that Muhammad Zahir Shah's regime had written in 1964. Afghans in Kabul (many of whom had recently returned from abroad) argued that this was necessary to prevent ethnic and regional division. They rejected alternative plans, labeling those who favored more regional autonomy and less presidential power as tools of the "warlords" or of ethnic blocs. The international community supported those Afghans in favor of centralization --  although the West's fears about ethnic destabilization reflected recent experiences in the Balkans more than Afghan reality. (Moreover, international advisers preferred dealing with a single central government rather than a series of local decision-makers, thus creating even more support for a strong executive power base.)

Like the U.S. constitution, Afghanistan's new constitution made no mention of formal political parties, and Karzai refused to allow candidates to run with any party affiliation or to let parliament organize itself by party. Karzai couched his opposition to political parties in the same rhetoric Zahir Shah had used in the 1960s: parties lead to national discord. But his decision only reinforced nonpolitical ties based on family, region, or ethnic affiliation. For lack of other alternatives, the Afghan parliament soon split into Pashtun and non-Pashtun blocs. Government ministries became ethnic enclaves, spoils that Karzai offered to ethnic leaders in exchange for their political support. At the same time, because Karzai unilaterally made all appointments, even down to the subprovincial level, palace patronage played a larger role than democratic politics in determining who governed at the local level. For regions that had grown accustomed to autonomy over the years, the arrival of Karzai's appointees, who abused their positions or favored one faction over another, created hostility to the central government. When they reemerged in southern Afghanistan in 2005, the Taliban drew on this resentment and on the belief that Karzai did not represent a national government so as much as a family network keen to reward political allies and punish rivals.


Afghanistan's various domestic problems have now become especially relevant for the United States and its allies as they look to withdraw their forces over the next few years, with the goal of transferring security responsibility to the Afghan government in 2014. A politically stable Afghanistan could accommodate a Taliban faction in government, whereas a politically unstable Afghanistan could not. Ethnic perceptions will play a role in determining what sort of country will emerge as U.S. and NATO forces begin to pull out. At the moment, non-Pashtun Afghans fear that Karzai will strike a deal with the Pashtun Taliban at their expense. They also believe that such a deal would not represent true power sharing but rather be a harbinger of the Taliban's return. Instead of fighting another bloody civil war, as Afghanistan's non-Pashtuns did in the 1990s, they might instead abandon the unitary state and secede, leaving the Taliban to struggle for power with other Pashtun factions in the south and east. Such a scenario would inevitably destabilize Pakistan's Pashtun-majority regions, making the vast, ungoverned territories that border Afghanistan even more anarchic and therefore fertile ground for various terrorist groups.

But it is not too late for Washington to avoid this scenario. First, the United States should make a greater effort to persuade Karzai to give legal recognition to political parties. At the moment, ethnic and regional networks owe their strength not to popular enthusiasm but to a simple lack of alternatives. Afghanistan's parliament would be far more effective and relevant if its members represented political platforms and agendas instead of geographically or ethnically defined constituencies. Such a reform would also reduce the defects in the existing winner-take-all balloting process, in which candidates run as individuals and the majority of parliamentarians receive less than 20 percent of the vote in their districts. The absence of political parties makes governance in Afghanistan more difficult and less legitimate while providing no offsetting benefit. Allowing any ideological group, including the Taliban, to legitimately and peacefully compete for a place at the national level would do more to ensure the stability of the Afghan government than any backroom deal that Karzai might make with the Taliban on his own.

Second, the United States should encourage the Afghan government to devolve power to provinces and districts so that citizens there can elect their own governors. Karzai currently makes all such appointments, but such authority is not mentioned in the 2004 constitution and could thus be changed through legislation. Once elected, governors should have the authority to raise local taxes to fund local services, a privilege that also now resides only with Karzai's administration in Kabul. Although it is true that Afghan governments since the late nineteenth century have resisted any devolution of governmental authority as too dangerous, these regimes were all run by kings and dictators. Afghanistan today is a nominal democracy, meaning that people in the provinces are less concerned that Kabul will make decisions they oppose. As one of the oldest democratic federal states in the world, the United States is in a unique position to make the case that in a diverse country such as Afghanistan, such a structure provides for more stability, not less.

Opening up provincial and district governorships to competition would provide the safest form of power sharing with the Taliban. Whereas non-Pashtun Afghans oppose granting the Taliban a role in the national government, they have few objections to former (or even current) Taliban members serving in districts or provinces where they have local support. Allowing the Taliban to serve in a democratic government would likely lead to beneficial fissures within the Taliban, since those who come to hold positions in local government would have less reason to remain loyal to the Taliban leadership based in Pakistan. Participating in a coalition government would put much different pressures on Taliban members from those they faced when they essentially ruled as dictators in the 1990s. The stated goal of the Taliban's central command -- seizing power nationwide -- would immediately clash with the interests of these local commanders turned politicians. Similarly, the need for these governors to deliver services and patronage to their own districts would increase their incentives to cooperate with those who could provide such aid: namely, the government in Kabul and its international allies. (An example of such a process already exists: some officials in the Karzai government are members of the Hezb-i-Islami, or Islamic Party, even though its leader, Hekmatyar, openly fights against Afghan and Western forces.)

To achieve even these modest goals, the United States and its allies must address a question that is still dangerously taboo: the status of 2014 as a transition date in Afghan politics. According to the Afghan constitution, a president can serve only two consecutive terms, meaning that Karzai must relinquish his office when his second term ends in 2014. No Afghan ruler has ever stepped down voluntarily, however, and Afghanistan is rife with speculation that Karzai intends to stay on regardless of the constitution. The United States and the rest of the international community should publicly announce their opposition to any extraconstitutional extension of the current presidency, if only to force Kabul's political class to begin considering the consequences of a future without Karzai -- and to convince Karzai himself that there is indeed such a future. Here, outside influence is especially important: most Afghans believe that without pressure from his patrons in the West, Karzai will not step down. Such moves from Washington and elsewhere need not be viewed as an attack on Karzai himself; after all, this is a constitutional issue and not a personal one, and Karzai has often argued that the Afghan constitution needs to be respected. Even the inkling that Kabul may have new leadership in 2014 would immediately open up Afghan politics to new ideas and personalities, particularly to the younger generation of Afghans who have so far been excluded from the political process. But the country would not have to wait until 2014 to begin to benefit from this change: after all, Karzai's current objections to political devolution and political parties might soften if he were to realize that someone else would soon wield the strong executive power that is currently his alone.

Along with structural changes such as political devolution and the allowance of political parties, the opening up of the political field in advance of 2014 offers the best possibility of creating a more stable and legitimate Afghan government. If Washington leaves the question of executive power unaddressed until 2014, however, then the much-heralded transition of responsibility to the Afghan government may founder over disputes about the government's legitimacy. Although ethnic and regional groups in Afghanistan have historically mobilized to fight when their interests were threatened, such reactions have been the product of pragmatism rather than any primordial hatred or nationalist ideology. The best way to avoid such conflict --  and, thus, to create a more stable Afghanistan -- is to address these interests before conflict arises, not after it starts.

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