The original plan for a post-Taliban Afghanistan called for rapid, transformational nation building. But such a vision no longer appears feasible, if it ever was. Many Americans are now skeptical that even a stable and acceptable outcome in Afghanistan is possible. They believe that Afghanistan has never been administered effectively and is simply ungovernable. Much of today's public opposition to the war centers on the widespread fear that whatever the military outcome, there is no Afghan political end state that is both acceptable and achievable at a reasonable cost.
The Obama administration appears to share the public's skepticism about the viability of a strong, centralized, Western-style government in Kabul. But it does not think such an ambitious outcome is necessary. As U.S. Secretary of Defense Robert Gates observed in 2009, Afghanistan does not need to become "a Central Asian Valhalla." Yet a Central Asian Somalia would presumably not suffice. Success in Afghanistan will thus mean arriving at an intermediate end state, somewhere between ideal and intolerable. The Obama administration must identify and describe what this end state might look like. Without clear limits on acceptable outcomes, the U.S. and NATO military campaign will be rudderless, as will any negotiation strategy for a settlement with the Taliban.
In fact, there is a range of acceptable and achievable outcomes for Afghanistan. None is perfect, and all would require sacrifice. But it is a mistake to assume that Afghanistan is somehow ungovernable or that any sacrifice would be wasted in the pursuit of an unachievable goal. Afghanistan's own history offers ample evidence of the kind of stable, decentralized governance that could meet today's demands without abandoning the country's current constitution. By learning from this history and from recent experience in Afghanistan and elsewhere, the United States can frame a workable definition of success in Afghanistan.
CONSENT OF THE GOVERNED
From the end of the Second Anglo-Afghan War in 1880 to the coup of Mohammad Daud Khan in 1973, Afghanistan underwent a relatively stable and gradual period of state building. Although the country was an absolute monarchy until 1964, Afghanistan's emirs, on the whole, needed the acquiescence of the population in order to govern. The central government lacked the strength and resources to exercise local control or provide public goods in many parts of the country. Instead, it ruled according to a series of bargains between the state and individual communities, exchanging relative autonomy for fealty and a modicum of order. Over time, as Kabul improved its capacity to offer services and to punish transgressors, this balance shifted, and local autonomy gradually eroded. But whenever this process went too quickly -- most notably in the 1920s under Amanullah Khan and in the 1970s under the Soviet-backed People's Democratic Party -- conflict in the periphery erupted and local power brokers challenged the central authority. The Soviet invasion in 1979 led to a fundamental breakdown of centralized authority and legitimacy, which resulted in the diffusion of political, economic, and military power across a number of ethnic and geographic groups. The era of dynastic control of the state by Pashtun elites is thus now over.
Although war, migration, and the emergence of regional strongmen have destabilized the Afghan countryside, local communities remain a fundamental source of Afghan identity and a critical base of governance and accountability. This is especially clear in the case of the local jirga or shura (community council). Traditionally, the community council was a place to solve problems and negotiate over common goods and burdens, with its more prominent members serving as liaisons to the central government. These bodies may differ in their power and representation, but they are still found today in virtually every community. This traditional and local base of legitimacy offers a potential foundation for stable governance in the future.
Washington, of course, would prefer to see Afghanistan -- much as it would like to see any country -- ruled in accordance with the will of the governed, its people prosperous, and the rights of its minorities and women respected. But the United States' two main security interests in Afghanistan that justify waging a war are much narrower: one, that terrorists who wish to strike the United States and its allies not use Afghanistan as their base, and two, that insurgent groups not use Afghanistan's territory to destabilize its neighbors, especially Pakistan.
There are many possible end states for Afghanistan, but only a few are compatible with these national security interests. Afghanistan could become a centralized democracy, a decentralized democracy, a regulated mix of democratic and nondemocratic territories, a partitioned collection of ministates, an anarchy, or a centralized dictatorship. The first and the last are unlikely; partition and anarchy are unacceptable. But decentralized democracy and internal mixed sovereignty are both feasible and acceptable.
THE FAILURE OF CENTRALIZATION
Since 2001, Hamid Karzai's government, with international support, has pursued the model of centralized democracy. As first envisaged in the 2001 Bonn agreement and then codified in the 2004 Afghan constitution, this approach places virtually all executive, legislative, and judicial authority in the national government. It has created one of the most centralized states in the world, at least on paper. The president appoints every significant official in the executive branch, from provincial governors down to midlevel functionaries serving at the subprovincial level. All security forces are national forces. Although there are provisions to elect provincial, district, municipal, and village councils, only provincial council elections have been held thus far. Kabul holds all policy, budgetary, and revenue-generating authority. In March 2010, Karzai approved a new governance policy that devolves some local administrative and fiscal authority to appointed officials and provides modest auditing and budgetary powers to elected subnational bodies. Still, the Afghan state retains a remarkably centralized blueprint.
Political figures close to Karzai pushed for such a highly centralized government against the wishes of many non-Pashtun minorities -- and despite Afghanistan's prior experience with failed, albeit nondemocratic, centralization efforts. From 1919 to 1929, for example, Amanullah Khan aspired to be Afghanistan's Kemal Atatürk, but his strategy ultimately led to serious rural upheaval, which ended his rule. The radical attempts at centralization under the Soviet-backed regimes that followed the 1978 coup helped spark the mujahideen resistance and led to years of civil war.
After the Taliban were removed from power, in 2001, strong Pashtun support, combined with fears of a return to the civil war of the 1990s, created a majority in favor of a centralizing constitution. But Afghan central governments have never enjoyed the legitimacy required by such an organizing principle. The last 30 years of upheaval and radical devolution of political, economic, and military authority have only made this problem worse. Put simply, the current model of Afghan governance is too radical a departure in a place where the central state has such limited legitimacy and capacity. To create a lasting peace that includes the country's main ethnic and sectarian groups -- as well as elements of the insurgency -- Afghanistan will require a more inclusive, flexible, and decentralized political arrangement.
Power sharing would be easier under a decentralized democracy, in which many responsibilities now held by Kabul would be delegated to the periphery. Some of these powers would surely include the authority to draft and enact budgets, to use traditional alternatives to centralized justice systems for some offenses, to elect or approve important officials who are now appointed by Kabul, and perhaps to collect local revenue and enforce local regulation.
Increasing local autonomy would make it easier to win over Afghans who distrust distant Kabul and would take advantage of a preexisting base of legitimacy and identity at the local level. The responsibility for foreign policy and internal security, however, would remain with the central government, which would prevent even the more autonomous territories from hosting international terrorist groups or supporting insurrection against the state.
A decentralized democracy along these lines should be an acceptable option for the United States. Its reliance on democracy and transparency is consistent with American values. Individual territories with the freedom to reflect local preferences may adopt social policies that many in the United States would see as regressive. But the opposite could also occur, with some places implementing more moderate laws than those favored by a conservative center. By promoting local acceptance of the central government, this option would remove much of the casus belli for the insurgency. And it would preserve a central state with the power and incentive to deny the use of Afghan soil for destabilizing Pakistan or planning attacks against the United States.
A decentralized democracy would comport with much of the post-Cold War experience with state building elsewhere. A range of postconflict states in Africa (Ethiopia and Sierra Leone), Europe, (Bosnia and Macedonia), the Middle East (Iraq and Lebanon), and Asia (East Timor and, tentatively, Nepal) have used some combination of consociationalism, federalism, and other forms of decentralized democratic power sharing. Although it is too early to make definitive claims of success, to date not one of these states has collapsed, relapsed into civil war, or hosted terrorists. And some, such as Bosnia and Ethiopia, have remained tolerably stable for over a decade. This is, of course, no guarantee that decentralized democracy would work in Afghanistan. But its track record elsewhere and its better fit with the country's natural distribution of power suggests that it offers a reasonable chance of balancing interests and adjudicating disputes in Afghanistan, too.
A decentralized democracy in Afghanistan would face three critical challenges. The first, of course, is the Taliban, who oppose democracy on principle and are likely to resist this approach as aggressively as they now resist centralized democracy. The second challenge is the limited administrative capacity of the Afghan state. Decentralization would distribute power among a larger number of officials; for a state such as Afghanistan, which has a limited pool of competent bureaucrats, this could exceed the country's current human capital and require a major expansion of training efforts. Third, the country's malign power brokers would likely resist such an option. A transparent electoral democracy would threaten their status, authority, and ability to profit from corruption and abuse.
Yet decentralized democracy could actually offer some important counterbalances in each of these areas. Hard fighting will be required to marginalize the Taliban under any democratic system, decentralized or not. The odds of success are much higher, however, when the population supports the government. Counterinsurgency can be described as a form of violent competition in governance; it is much easier to win when the form of government offered is closer to the natural preference of the governed. And if the Taliban come to see their military prospects as limited, a decentralized system might entice some of their members to reconcile with the government in the hope of securing a meaningful local role in areas where their support is strongest.
It will not be easy to combat high-level corruption or to improve administrative capacity. But a transparent system in which locals make most decisions would allow Afghanistan's traditional community leaders to police the use of power and public funds. A faraway national ministry in Kabul is beyond the oversight of a village or district shura. In contrast, local councils can see how officials are spending money and can take issue with uses they find objectionable. Decentralization may also improve the Afghan government's basic competence by allowing local officials to focus on smaller, more local issues. For example, the most widely hailed development program in Afghanistan in the last eight years has been the National Solidarity Program, under which the central government provides grants to democratically elected community councils for local development projects. The NSP was designed at the national level but is administered locally. To date, it has been fiscally efficient and effective, reaching more than 20,000 villages.
Although decentralized democracy offers no easy guarantee of success, it has much better odds of success than a centralized model. But it would not come cheaply: the United States would have to wage a sustained counterinsurgency campaign, provide major administrative assistance to the Afghan government, and conduct vigorous anticorruption measures.
A MIXED BAG
Mixed sovereignty is an even more decentralized model. Much like decentralized democracy, this approach would take many powers that are now held in Kabul and delegate them to the provincial or district level. But mixed sovereignty would go one step further, granting local authorities the additional power to rule without transparency or elections if they so chose -- as long as they did not cross three "redlines" imposed by the center.
The first redline would forbid local authorities from allowing their territories to be used in ways that violated the foreign policy of the state -- namely, by hosting terrorist or insurgent camps. The second would bar local administrations from infringing on the rights of neighboring provinces or districts by, for example, seizing assets or diverting water resources. The third would prevent officials from engaging in large-scale theft, narcotics trafficking, or the exploitation of state-owned natural resources.
Beyond these limited restrictions, local authorities could run their localities as they saw fit, with the freedom to ignore the will of the governed or engage in moderate-scale corruption. The central government in Kabul would retain total control over foreign policy and the ability to make war and enforce narcotics, customs, and mining laws and limited authority over interprovincial commerce. Under such an arrangement, sovereignty is mixed to a much greater degree than in the other possible systems, with many -- but not all -- of the ordinary powers of sovereign government delegated to the provincial or district level.
The mixed-sovereignty model would signal a more serious break with the direction of Afghan state building as it was conceived in 2001 than would decentralized democracy. But it would also be a partial acknowledgment of the de facto arrangements that have taken shape since 2001. Many of the governors and other local officials appointed by Karzai have ruled not by virtue of a legal mandate from Kabul but rather through their own local security and economic power bases, which operate outside the law but with the tacit acceptance of Kabul. In provinces such as Balkh (under Governor Atta Mohammad Noor) and Nangarhar (under Governor Gul Agha Sherzai), this has led to relative peace and a drastic reduction of poppy cultivation. Such warlords have settled into a stable equilibrium in which they profit from the theft of customs duties and state property but maintain order and keep their predation within limits so as to avert a mutually costly crackdown by Kabul.
In other areas, however, strongmen have caused instability. In Helmand, for example, several years of corrupt rule by Sher Mohammad Akhundzada alienated significant groups in the province and sent poppy cultivation soaring, fueling the insurgency. Even in Afghanistan's relatively stable north, the rule of warlords has led to ethnic violence and criminal excess. To ensure stability, mixed sovereignty cannot amount to partition under local strongmen who rule with impunity in private fiefdoms. Redline restrictions that forbid the sort of excesses that fuel insurgency are thus essential.
Mixed sovereignty has some important advantages: it is less dependent on the rapid development of state institutions and offers a closer fit with the realities of Afghanistan. Restricting the central government's involvement in local issues to a limited -- but aggressively enforced -- set of redlines could encourage the country's power brokers to moderate their excesses, which now drive many toward the Taliban. At the same time, a mixed-sovereignty system would depend less on transparency and efficiency, thus requiring less international mentoring, oversight, and assistance. Local autonomy would create incentives for Taliban members to participate in reconciliation negotiations, since a more purely democratic option would subject them to electoral sanction.
However, mixed sovereignty also carries risks and disadvantages that make it less consistent with U.S. interests than either centralized or decentralized democracy. First, governors would be free to adopt regressive social policies and abuse human rights. This would represent a retreat from nearly nine years of U.S. promises of democracy, the rule of law, and basic rights for women and minorities, with costs to innocent Afghans and the prestige of the United States.
Corruption would also be prevalent -- indeed, for prospective governors, the opportunity for graft would be an essential part of the system's appeal. The Afghan government would have to contain the scale and scope of this corruption, lest official acceptance of abuse renewed support for the insurgency. To prevent this, Kabul would have to rein in the worst of today's excesses -- if mixed sovereignty is merely a gloss for the status quo, it will fail. At the same time, the Afghan state would have to crack down on the narcotics trade, which if left unchecked could dwarf the revenues provided by foreign aid and make such aid a less convincing incentive for compliance with the center. The central government would have to strike a bargain with the country's power brokers, requiring them to refrain from large-scale abuses in exchange for tolerance of moderate local corruption and a share of foreign assistance. Even this kind of bargain, however, would probably be resisted by the country's strongmen, who have grown used to operating without restraint. Thus, mixed sovereignty would not free Kabul from the need to confront local power centers, and even this limited confrontation could be costly and difficult.
Under this style of governance, there would be a potential threat of instability as powerful governors periodically tested the waters to see what they could get away with. The central government would presumably need to carry out periodic enforcement actions, including violent ones.
Mixed sovereignty is thus not ideal, but it could be viable and meet U.S. security requirements if Washington and Kabul were willing to fulfill their roles as limited but important enforcers. The model offers the central government two means of imposing the essential redlines. The first is the threat of punitive military action ordered by Kabul. This would require security forces that have the capability to inflict serious costs on violators. (They need not have a monopoly on violence, but a meaningful national military of some sort is necessary.) The other enforcement mechanism is Kabul's control over foreign aid and its ability to direct aid to some provinces but not others.
Washington would not be powerless, either -- it would retain its influence through the disbursement of foreign aid and its deep engagement with the Afghan National Security Forces. In order to maintain Afghanistan's internal balance of power, the United States and its NATO allies would need to pay constant attention. Otherwise, the country could slip into unrestrained warlordism and civil war. A workable mixed-sovereignty model is not a recipe for Western disengagement: it would require not only continued aid flows but also sustained political and military engagement. Regional diplomacy would be particularly important. To keep Afghanistan from becoming a magnet for foreign interference and a source of regional instability, the United States would have to ensure that the country was embedded in a regional security framework. Such a framework would facilitate aid flows and discourage intervention by Afghanistan's neighbors.
As with decentralized democracy, internal mixed sovereignty has produced tolerable outcomes in the developing world. Afghanistan itself was governed under a similar model for much of the twentieth century: Muhammad Nadir Shah and his son Muhammad Zahir Shah ruled for five decades as nominally absolute monarchs, but with limited state bureaucracy and a certain degree of autonomy for the periphery. The rule of law was generally administered locally, and some Pashtun tribes in the south and the east were exempted from military service. Nevertheless, a national army and a national police force remained ready to enforce a few key royal prerogatives. The government earned revenue not from internal taxation but from foreign trade, foreign aid (starting in the late 1950s), and the sale of natural gas to the Soviet Union (beginning in the late 1960s). Over time, as the government's capacity and resources increased, it was able to extend its writ, trying criminals in state courts, regulating the price of staple goods, and bringing community land under its authority.
There are also external parallels. After the end of the Nigerian Civil War in 1970, Nigeria had a weak federal government and a strong regional system, in which individual governors were free to organize local administration as they wished. Even today, the country retains some traits of internal mixed sovereignty. States in the Muslim north have sharia law, whereas others use secular judicial systems. The central government intervenes selectively to suppress unrest, such as in the Delta region. Although there are signs that Nigeria may now be deteriorating, for most of the last 40 years it has functioned tolerably.
THE UNACCEPTABLE OTHERS
Many other outcomes for Afghanistan are possible -- but would fail to meet core U.S. security requirements. The country could, for example, split up in a form of either de facto or de jure partition. The most likely such split would divide the Pashtun south from the largely Tajik, Uzbek, and Hazara north and west. Such a result could come about if a reconciliation deal with the Taliban granted the group too much leeway in the country's south, its historical power base. Any outcome that leaves the Taliban relatively free to operate in the south could create safe havens for cross-border terrorism and insurgency, similar to the use of Iraqi Kurdistan by the Kurdistan Workers' Party, or PKK, or the use of Congolese border havens by Hutu guerillas. Partition would also set the stage for regional proxy battles and internal competition for control of Kabul and key border areas.
If the Karzai government collapses, Afghanistan could break down into the kind of anarchy and atomized civil warfare of the 1990s. Such a state would resemble the one that was taken over by the Taliban in the 1990s, or present-day Somalia, where lawlessness has created an opening for al Shabab, a violent, al Qaeda-supported Islamist movement -- with obvious consequences for U.S. interests.
Lastly, Afghanistan could become a centralized dictatorship, although this is hard to imagine. A single strongman is unlikely to be able to consolidate power in post-Taliban Afghanistan, where political, military, and economic might is dispersed among numerous power brokers. In this environment, any prospective dictator -- whether pro- or anti-Western -- would find it very difficult to prevent the country from descending into civil war. A coup d'état or other antidemocratic power grab (amending the constitution, for example, to allow for a president for life) is entirely possible but unlikely to yield stability in its wake.
SALVAGING THE GOOD
Afghanistan has been a failing experiment in centralized democracy, heading toward de facto partition, with Taliban control in some areas and unstable, ill-regulated strongman governance in many others. This trend can be reversed. But clinging to the original, centralized model will not help. Centralized governance matches neither the real internal distribution of power in Afghanistan nor local notions of legitimacy. There can be no effective military solution if the intended political goal is so badly misaligned with the country's underlying social and political framework.
To its credit, the Obama administration appears to have recognized that centralized democracy is a bridge too far for Afghanistan. Current policy is moving toward decentralization -- the question is how far this should go and whether Afghan and U.S. officials can manage the transition successfully.
This shift toward decentralization can work, although it is no panacea. A system of either decentralized democracy or internal mixed sovereignty would have its drawbacks, and each would involve sacrifice and risk. In Afghanistan -- as in most places -- the more optimal a system of governance, the longer and harder the fight to get it. The question of whether to strive for the preferable outcome of decentralized democracy or to accept the less appealing alternative of internal mixed sovereignty will largely be determined by the efforts and sacrifices the United States and its partners are willing to undertake. Yet for all their drawbacks, either approach would meet core U.S. national security requirements if properly implemented. And either model is more achievable than today's goal of centralized democracy.
Moreover, a decentralized democracy would not require the Afghan government to abandon or amend the existing constitution. The 2004 constitution is flexible enough to allow many powers to be devolved through legislation, as demonstrated somewhat by the new subnational governance policy, which provides limited administrative and budgetary authority to local officials. A mixed-sovereignty model would clash with the spirit and letter of the 2004 constitution, but such a system would likely evolve on a de facto basis, averting the need for a new constitution in the near term.
Afghanistan is not ungovernable. There are feasible options for acceptable end states that would meet core U.S. security interests and place the country on a path toward tolerable stability. The United States will have to step back from its ambitious but unrealistic project to create a strong, centralized Afghan state. If it does, then a range of power-sharing models could balance the needs of Afghanistan's internal factions and constituencies in ways that today's design cannot, while ensuring that Afghanistan does not again become a base for terrorists. In war, as in so many other things, the perfect can be the enemy of the good. The perfect is probably not achievable in Afghanistan -- but the acceptable can still be salvaged.