The relationship of the central government of Afghanistan to the other units of government is in many ways a proxy for the relationship of state to society. It would not be so if the state were more institutionalized and in control of the territory and population of the country. But the current situation, where the direct administrative control of the government is largely limited to the capital city and environs, and in which the government relies on international support (“foreign” support to its opponents) to exercise that control, has precedents in other eras of Afghan history.
The Bonn agreement, which provides the legal framework for the current government, reinstated the administrative provisions of the Constitution of 1964. Under that constitution Afghanistan is a unitary state administered according to the “principle of centralization.” Laws enacted by succeeding governments have divided that unitary state into thirty-two provinces (wilayat), three of which were established during the past twenty-five years of conflict. The number of districts (wuluswalis) into which these provinces are subdivided has been a matter of controversy, as local authorities claim to have established new districts in a number of locations that the ministry of internal affairs and other authorities do not recognize. Some wuluswalis were further divided into alaqadaris (subdistricts), but these became recognized as wuluswalis, apparently during the period of rule by President Rabbani. According to the Minister of Internal Affairs, Ali Ahmad Jalali, the country had 216 legally established wuluswalis before the start of the conflict; it now has 354. The Independent Commission for Convening the Emergency Loya Jirga identified 383 claimed wuluswalis in the country. The system also includes municipalities and rural municipalities.
The governors, district commissioners, and other officials of this structure are, legally speaking, representatives of the centralized unitary state in their areas, appointed directly by the minister of internal affairs and the president. The provincial and local representatives of other ministries (education, agriculture, water and power, etc.) report directly to their ministries in Kabul. The governor has only a loose coordinating role and does not have executive authority over the representatives of ministries other than interior. There is no provincial budget or system of block grants. Provisions of past constitutions providing for provincial or district councils to participate in decisions about the execution of policy were never implemented. The judicial system was similarly centralized, as the same law applied uniformly in all provinces, and all courts came under the direct authority of the Supreme Court in Kabul.
Even before the past quarter century of conflict, this highly centralized system had very limited reach. The lowest level of territorial administration at which the state was present was generally the district, or sometimes the subdistrict. The widely reported institutions of local self-government and dispute resolution (jirga, shura) had no legal existence; nor did the constitution or legal code mention or recognize the widely used systems of customary law. The villages, in which most people lived, appointed representatives (malik, arbab) who represented them in dealings with the wuluswal or governor. Government policing did not extend outside of the district center. The base of taxation was extremely small: the government’s revenue remained under 10 percent of GDP and came almost entirely from taxes on foreign trade, state monopolies (petroleum, tobacco), sale of natural gas, and foreign aid. Hence the state apparatus barely penetrated the country’s economy and in particular hardly taxed at all the predominant activities of agriculture and pasturage, with the exception of a small set of commercial export crops. Social norms dictated that disputes should be settled as much as possible outside official institutions. Even those dissatisfied with the outcome of a jirga or shura might not want to disrupt local social relations or risk the pressure for bribes that came with resort to the official system. Hence the system whose legal framework is now in effect was a highly centralized but weak state, with very limited penetration into the society and scope for policy making or implementation.
The reality of the exercise of power today is quite different. Although the legal administrative subnational unit is the province, of which there are thirty-two, the territorial unit over which powerful figures actually exercise subnational power is the region or zone (hawza), of which there are about seven in addition to Kabul. These zones correspond to the old provinces of Afghanistan, which in turn were related to tribal or ethnic identities that have reemerged as transformed units of political identity and mobilization as a result of the war. The monarchy and its successors developed the current framework of territorial administration precisely to divide, weaken, and overpower the tribal and regional forces of these ancient zones, and the reemergence of these units results from the breakdown and weakening of this always fragile superstructure of control.
In analyzing the politics of relations between the center and other levels of government in Afghanistan, it is important to bear in mind the distinction of state and administration. Elements of the administrative structure of Afghanistan have remained relatively stable for centuries. In his memoirs, the Baburnama, Zahiruddin Muhammad Babur, who became the first Mughal emperor of India, recounts his journey of conquest and plunder across much of the territory of today’s Afghanistan. Throughout his travels he describes the districts through which he passes and characterizes the main sources of revenue in each one. Though Babur led one of a number of roving warrior bands seeking to build states on this territory and neighboring ones, the instability of the state structures contrasted with the stability of the administrative ones, which had been erected during previous centuries of imperial rule. The combination of these districts into provinces and of provinces into emirates, khanates, kingdoms, and empires varied, but the administrative substructure constituted a relatively stable reality for ruler and subject alike.
No one saw the construction of states as a means to express the identity or to protect the rights of those who lived in them. States were organizations through which conquerors exercised control over population and territories. States could create security that enabled merchants to increase wealth, in return for which they paid tribute to the state. The legitimacy of a state and its ruler depended on the degree of justice the ruler delivered, as defined by Islam, but the provision of justice was a duty of the ruler, not a right of the subjects, who were not citizens.
The concept that control of the state constitutes war booty belonging to the victor appears to remain embedded in much of the politics of the region. In today’s Afghanistan it is expressed by those among the “mujahidin,” especially the Panjshiris in Shura-yi Nazar, who say that they deserve to predominate in the government because they resisted the Soviets, the Communists, and the Taliban and finally prevailed. It explains why they regard sharing power, in however limited a way, as a significant concession. The prevalence of this view should serve to caution observers that those who protest Panjshiri domination may not be democrats, but contenders for capture of that same booty—the state—for other groups.
The term Afghan was originally an ethnonym for the people today known as Pashtuns. Ahmad Shah Durrani, the ruler who in 1747 founded the empire from which today’s Afghanistan descends, held the title of “King (Padishah) of the Afghans,” not “King of Afghanistan.” Afghanistan at that time denoted not a state but a territory, the area now known as the tribal region shared by Afghanistan and Pakistan.
The state founded in 1747 was a tribal conquest empire. According to official accounts, the king was elected at a jirga of the tribes of Kandahar, primarily belonging to the Abdali confederation (wulus), which was then renamed “Durrani,” after a royal epithet. This jirga should not be confused with the loya jirgas established by later rulers as a legitimation device. At that time, the tribes constituted the military force of the empire, and their agreement on a matter was self-enforcing. Led by Ahmad Shah, the tribes formed alliances with other groups and conquered richer revenue-producing areas such as Kabul, Herat, Punjab, Kashmir, and Turkistan (today’s north Afghanistan). The state apparatus erected by Ahmad Shah and his successors was a mechanism by which the Pashtun tribes under his command ruled over other territories. It extracted revenues from conquered areas and remitted them to Kandahar (to Kabul, after 1775, when it became the capital). It did not tax the ruling tribes. The ruler constantly struggled to gain autonomy from the tribes by building an army directly dependent on him, often recruited from non-Pashtun groups, and paid for by royal revenues.
This history forms one level of the consciousness of the people of Afghanistan about the role of the state. As recently as 1959, Kandaharis rioted when the government tried to collect land tax from the province. Some Pashtuns, and particularly Kandaharis, view the Afghan state as a mechanism through which a Pashtun-dominated elite should rule over others. Many non-Pashtuns have viewed it the same way and therefore want to transform it. The calls for “federalism” from some quarters largely express this protest.
It would be mistaken to regard the Pashtun character of the state as a manifestation of ethnic politics in the modem sense. The state never attempted to incorporate all Pashtuns into the ruling group, even symbolically. Rulership always resided in a particular lineage: for the first decades after 1747 in the Saddozai lineage of the Popalzai tribe, and thereafter in the Muhammadzai lineage of the Barakzai tribe. These lineages competed, as did various family and clan groups within each lineage. In these battles ethnic considerations were largely irrelevant. Muhammadzais fought Muhammadzais allied with Uzbeks, Tajiks, Ghilzai Pashtuns, or anyone else whose allegiance could be won or bought. The state was never exclusively Pashtun in its composition or staffing, even at high levels, and the idea of ethnic purity would have seemed a nonsensical and useless limitation of the ruler’s freedom of maneuver. The state did not attempt to legitimate itself in the name of Pashtun nationalism, culture, or language until the 1930s. Despite the narrow base of rulership, legitimation was expressed through the universal values of Islam, in particular by the role of the king as defender and supporter of the faith and by establishing courts of sharia at least in the major towns.
The major provinces corresponded to regional kingdoms, each consisting of a market center and its hinterlands: Herat, Kandahar, Kabul, and Turkistan (centered on Balkh, then Mazar-i Sharif). The king (amir during 1826–27) generally appointed his brothers or sons as governors to reduce the risk of revolt, but it nonetheless occurred, as tax farmers converted themselves into contenders for rule.
Amir Abd al-Rahman Khan (reigned 1880–1901) converted this military feudal state into one with centralizing absolutist aspirations. He could do so because of the financial and military subsidies he received from the British Empire, which enabled him to break the military power of the tribes and peoples of all regions of Afghanistan, collect taxes from them, and establish a centralized military, administrative, and judicial system. He brought nearly all the Muhammadzais to Kabul, separating them from their tribal base, and named them “Partners of the State” (Sharik al-Dawlat). The state provided them with subsidies and appointed them disproportionately to offices. They thus became a court nobility dependent on the patronage of the amir rather than an independent tribal military power.
Amir Abd al-Rahman Khan established the basic institutional framework of the Afghan state that has now been largely destroyed and that the current government of Afghanistan is trying to reestablish. He kept his family at court and appointed professional administrators and generals rather than family members. Using his coercive apparatus, he penetrated the economy more deeply than subsequent Afghan states did, collecting direct taxes on agricultural produce and land, in addition to the subsidies from British India. He legitimated his absolutist rule through Islam, claiming to be the imam of Afghanistan, who defended the realm of Islam through jihad and enforced sharia, to whom the people therefore owed obedience. He focused rule and legitimation on his person and dynasty, not, as before, on the entire Muhammadzai clan, whose elders agreed not to contest his dynasty’s authority.
He conquered the autonomous regions of Hazarajat and Nuristan, converting the latter to Sunni Islam from paganism and enslaving many of the Shia Hazaras. He mobilized Ghilzai and Durrani Pashtuns against each other and exiled many of their khans and elders in an attempt to break down tribal power, but not before he used Pashtun tribal armies to conquer the Hazaras. By his own estimate he crushed forty rebellions and won eight civil wars. He erected piles of skulls on many roadsides and kept the executioners and torturers busy, but he died in his bed and passed his rule uncontested to his son, the first—and so far the last—time that rulership was transferred peacefully and legally in Afghanistan. He did so by suppressing the military-political organizations based in Afghanistan’s regions and subjecting the population instead to a hierarchical, centralized, absolutist system of rule.
This version of the centralized system continued until 1928, when a revolt based in the Tajik peasantry north of Kabul ousted King Amanullah Khan, grandson of Amir Abd al-Rahman Khan. Amanullah, Afghanistan’s autocratic reformer, had given the country its first constitution in 1923, its first budget, and first development program. Amanullah Khan’s government divided the four large provinces into ten smaller ones, but it also appointed military generals as governors-general (rais-i tanzimiyya) over regional groupings of provinces. [In 2011, with the appointment of presidential confidante Asadullah Khalid as “security coordinator” for the south after the assassination of Ahmad Wali Karzai, President Karzai appeared to be reviving that system.] Reduction of the size of provinces aimed to increase the leverage of the administration and break up the society into more manageable and smaller units, while the governors-general served as trouble shooters, assuring the governors’ loyalty to Kabul and adherence to rules.
The constitution of 1923 mentioned “decentralization of power” (ghair-i markazi budan-i qudrat) as one of the principles of the administration, but this seemed to refer to devolution of power to the officer in the field, like the British Indian Civil Service, rather than to the people themselves. Amanullah’s attempt to tax the peasantry and the border tribes to finance a program of Westernization and modernization provoked a revolt where opposition to state penetration invoked the symbols of Islam.
The Muhammadzai dynasty (that of Nadir Shah) that ousted the Tajik Amir Habibullah after nine months took a different tack. Again, the state and capital (which was looted) had been captured as booty by a coalition of Pashtun tribes, this time mostly from eastern Afghanistan, since Nadir Shah had entered Afghanistan from British India. These tribes received exemptions from conscription and taxation. They engaged in punitive raids against the Tajik areas north of Kabul, including Panjshir, kidnapping women among other forms of loot. The memories of these raids are alive today and affect perceptions of contemporary politics.
Nadir Shah and his brothers, who acted as regents for Nadir’s son Zahir Shah after the former’s assassination in 1933 by a partisan of Amanullah Khan, decided to encapsulate rather than penetrate and transform rural society. Until the mid-1950s, they tried to develop state-supported commercial agriculture on newly irrigated lands settled by Pashtuns in north Afghanistan. Cotton was the centerpiece of their strategy, as it was in Soviet Central Asia directly north of Afghanistan. The income from these new crops supported a very modest expansion of state institutions, while direct taxation of the peasantry declined. The administration continued to be centralized, as before, but it ceased its ambitious efforts to penetrate and control the society.
From the mid-1950s, the availability of foreign aid from the Cold War rivals enabled the Afghan government under Prime Minister Daoud Khan to pursue a variation on the strategy of encapsulation. It nationalized the Afghan development bank (Bank-i Milli) and built a new army, educational system, roads, dams, and other development projects with foreign aid. The state, the capital city, and the class of state-educated people who ran the state and lived in the capital greatly expanded as a result, but the direct tax base of the state continued to shrink. As the state expanded, its penetration into rural society under the level of the district actually contracted.
At the same time, a new urban, educated class emerged. New Democracy during 1963–1973 was designed to give them more of a voice in the state they served, but the failure to legalize parties drove many of them underground. The parliament was dominated by local notables and ulama, whose main legislative success was the further reduction of taxation of land and flocks. Hence the professional intelligentsia became increasingly frustrated by the power of rural elites who had been weakened but not defeated by the state. During this period the ten provinces of the 1950s were divided into twenty-nine.
The coup d’état by Daoud Khan in 1973 marked the first time that these elites effected a change in power, in their role as army officers and bureaucrats. The communist coup in 1978 was an even more direct seizure of power by one faction of the intelligentsia, since it did not have the cover of being led by a member of the royal family, as in 1973.
The attempt by the communists to use the state to transform the society with Soviet backing quickly exposed the weakness of that apparatus. Revolt broke out quickly, but the pattern of revolt is quite significant. The first two regions to escape from state control almost completely were Hazarajat and Nuristan, the last two regions conquered by the Afghan state. Hazarajat became more or less united initially under a Council of Unity (Shura-yi Ittifaq), while Nuristan eventually split into two parts, the Islamic State of Nuristan under Mawlawi Afzal and a less consolidated area.
Although both of these units incorporated districts that remained more or less unchanged, their somewhat amorphous boundaries did not coincide with the provincial ones demarcated by the Afghan state. They corresponded instead to ethnic identities that were identified with geographical homelands. Neither of these units was stable (Hazarajat in particular was torn apart by conflict between Iranian-supported groups and more traditional ones), and neither of them sought independence from Afghanistan. The importance of Nuristanis to the Afghan state elite and of Hazaras in the cities of Kabul and Mazar-i Sharif precluded any such demands. Nonetheless, the establishment of de facto autonomous authorities spanning bureaucratically defined provincial borders created the base for challenging the relations between state and society in Afghanistan. However imperfectly, the administration in these territories represented the identity of those living there, not control by an elite in Kabul.
In 1991, the Republic of Afghanistan under Najibullah recognized the de facto autonomy of Nuristan by creating a new province of that name from districts of Kunar and Laghman. In 1990 Najibullah created the new province of Sar-i Pul from predominantly Hazara districts of Balkh, Samangan, and Jauzjan. In 2003, the Hazara leader (and minister of planning) Muhammad Muhaqqiq was developing plans to separate predominantly Hazara districts of Ghazni, Uruzgan, and Ghor, presumably preparatory to the creation of a predominantly Hazara Central Zone. He has taken these proposals to President Karzai, who initially refused, as such changes would be quite provocative to Pashtuns at this time. [Since this paper for the World Bank was written in 2003, the predominantly Hazara district of Daikundi has been separated from Uruzgan and made a province in its own right.]
Hazarajat and Nuristan are both mountainous areas with few roads or resources and of relatively little strategic value. Hence the communist regime and the Soviet military more or less left them alone. Other areas of the country, however, they managed to keep divided and separated, as long as the Soviet troops were present.
The administration began to collapse in 1979, before the arrival of Soviet troops. Starting with the mutiny of the Herat garrison in March 1979, led by captains Ismail Khan and Alauddin Khan, nearly every major military garrison and many others erupted in revolt, refusing to enforce the regime’s programs and repression. Much of the army dispersed or defected to the growing mujahidin revolt, though that revolt proved unable to absorb professional military units or officers. The resultant growth of insecurity and of predation along the highways—a combination of criminal and political activities that were often difficult to distinguish—made the government unable to deliver salaries or operating funds to some provinces and many districts. Many governors and wuluswals lived in Kabul, unable to take up their posts. Local army units often reached accommodation with the local mujahidin: the former stayed in their garrison, while the mujahidin ran the locality, to the extent that anyone did. In addition to the Soviet bombing and indiscriminate offensives, the insecurity and factional fighting that resulted from the collapse of administration and growth of fragmented guerrilla forces also contributed to the refugee exodus.
During this period the Soviets and the Afghan military established a system of zones (hawza), each commanded by a general. Though these generals had no formal role in the civil administration, they developed roles similar to that of the governors-general. It is not clear at this point how this zonal division related to the emergence of regional militia formations, but both reflected a geographical reconfiguration of power.
As the Soviet troop withdrawal proceeded, and especially after it was completed, mujahidin and militia commanders began to develop regional political military formations like those in Hazarajat and Nuristan in different parts of the country. The government, unable to rely on the Soviet troop presence anymore, also used freshly printed money to pay off commanders, whose “tribal” (qawmi) militias also exercised regional influence and increasingly came to resemble “mujahidin” units. The major efforts were as follows:
• In 1986 Ahmad Shah Massoud founded the Supervisory Council of the North (Shura-yi Nazar-i Shamali), which originally functioned as a coordinating body for forces mainly affiliated with Jamiat-i Islami and mainly Tajik in northeast Afghanistan. SCN eventually exercised influence over parts of Kapisa, Parwan, Kabul, Kunduz, Baghlan, Balkh, Takhar, and Badakhshan provinces. After the Soviet withdrawal Massoud captured the town of Taluqan, capital of Takhar province, which served as the de facto capital of Shura-yi Nazar’s administration in the northeast. This quasi-state included its own armed forces (the core of those now commanded by Defense Minister Muhammed Qasim Fahim) and departments of education, finance, interior, and others. It was mainly funded by foreign aid, the emerald trade, humanitarian assistance, and, perhaps, the drug trade.
• Ismail Khan emerged as the predominant commander of Western Afghanistan. In 1988 he called a meeting of internal commanders from all over Afghanistan, though most of those that came were from the West. Massoud did not attend, which became a point of contention. Through this and similar gatherings Ismail Khan consolidated his influence in the area, though he could not emerge openly as a regional commander until the fall of the Herat garrison in 1992.
• Abdul Rashid Dostum advanced from a security guard at the natural gas fields in his home province of Jauzjan to the commander of the most powerful of the militias of Najibullah’s regime, recruiting largely from his own Uzbek ethnic group. This militia defended the security of the road from Hairatan on the Afghan-Soviet (now Uzbekistan) border to the Salang pass, where another militia, composed of Ismaili Hazaras, took over. This road constituted the regime’s main supply route and this militia consequently received a high level of support.
• In Kandahar, the commanders were unable to form a common military front, but the ulama established an Islamic court of Baluchistan and Southwest Afghanistan that dealt with the affairs of Afghans in both Pakistan and Afghanistan. In many respects this court rather than the commanders became the most influential body in southwest Afghanistan. Many of its judges later became officials of the Taliban.
Others had tried and failed to carry out such schemes. The late Abdul Haq, a commander from the prominent Arsala family of the Jabbarkhel clan of the Ahmadzai tribe in Nangarhar province, brother of the late Vice-President Hajji Abdul Qadir, and of the present governor of Nangarhar, Hajji Din Muhammad, tried to extend his influence from Nangarhar to Kabul and form a common front in Eastern Afghanistan. His refusal to destroy the country’s essential infrastructure (electric pylons in particular), however, led Pakistan to cut off his assistance. The southeastern provinces of the Pashtun tribal zone remained largely fragmented, though Jalaluddin Haqqani of Paktia presided over a loose tribal shura. He later became an important Taliban commander.
In an attempt to consolidate some of these efforts, major commanders, with support from Pakistan and the United States, held a meeting of a “national commanders’ shura” (NCS) in early 1990. In 1989, as the Soviets were leaving Afghanistan, the United States, Pakistan, and Saudi Arabia had supported a gathering in Rawalpindi that chose an Interim Islamic Government of Afghanistan. Commanders inside Afghanistan, though affiliated to the same parties as those that chose the IIGA, largely opposed the move and argued instead that those leading the fighting inside Afghanistan should build a new government. The NCS presented itself as no more than a means of military coordination, but it also embodied a potential state-building project. The shura opposed the attempt to establish the IIGA in Afghanistan through attacks by Pakistan-based Afghan militias, and instead called for the establishment of nine zones, each led by a prominent commander, that would capture provincial outposts and eventually move against Kabul.
The NCS thus proposed a decentralized strategy instead of the centralized, top-down one supported by Pakistan. The zones would each have included several provinces or portions thereof and would have been led by commanders originating in that region. The model of creating a new Afghan state from locally led and organized zones seemed to presage a federation. When one author (Rubin) called the NCS resolution a model for a federal state, however, he learned how sensitive Afghans were to this term. Rubin had used this term in a paper he wrote for the Carter Center in January 1992, and which he sent to Muhammad Eshaq, then editor of AfghaNEWS, a Peshawar weekly newspaper representing Ahmad Shah Massoud’s views. Today Eshaq is the director of Radio Television Afghanistan in the Ministry of Information and Culture. Eshaq faxed a lengthy refutation of the charge that this model represented federalism, which he said would be a formula for the division of the country. The NCS was solely an organizational model for military operations, he claimed.
When Najibullah’s government split and fell, to be succeeded by the Islamic State of Afghanistan (ISA), the country attained the highest degree of fragmentation. Massoud emerged as the predominant leader in Kabul, but only through constant battles that destroyed the capital, first with Gulbuddin Hikmatyar’s Hizb-i Islami and then with his former allies. He continued to rule much of northeast Afghanistan through SCN. Dostum formed an alliance with other regime militias and some mostly Uzbek mujahidin groups to form the National Islamic Movement of Afghanistan (Junbish, NIMA), which predominated in the north, though it fought with elements of SCN over control of Mazar-i Sharif and Kunduz, both of which controlled important bazaars and trade routes. Central Afghanistan was dominated by Hizb-i Wahdat. Ismail Khan dominated the west as “Amir of Herat,” and collected the customs revenue from the burgeoning trade with Iran. A shura in Kandahar never pacified the area, and opium growing predominated in the southwest. The southeast was even more fragmented. A shura in Jalalabad exercised a kind of loose dominance over the country’s second-largest opium-growing region. Hikmatyar continued to use the east and southeast as staging areas for his attacks on Massoud, though he controlled little territory himself. Thus the country was divided into about eight zones, including Kabul and environs, each reflecting a particular ethnic mix and leadership. Kabul’s main income was printing money, as the zones now retained their own customs revenue. It sometimes paid the salaries of officials in some areas, but the continuous issuance of new money had reduced the value of these salaries to almost nothing. The lack of central control over the provinces was also instrumental to the increased freedom of action of the Arab militants in the country, mostly in the eastern and southeast regions. It also made it possible for Usama Bin Laden to return to Afghanistan when he was expelled from Sudan in May 1996.
Although power was fragmented, UN agencies and NGOs attempted to deliver services to people, largely in rural areas. To do so they had to develop a stable method of interaction with the counterparts. The Swedish Committee for Afghanistan, which had pioneered cross-border assistance, had always insisted that local commanders settle their differences before it would deliver aid to the locality. When USAID and the CIA began to deliver cross-border “humanitarian” assistance from Pakistan to Afghanistan, they tried to create a centralized bureaucracy from the staff of the jihadi parties in Pakistan. Most NGOs and agencies, however, developed the SCA model further by telling villages that, in order to receive aid programs, they had to “revive” their traditional “shuras.” Several programs of UNDP and UN Habitat focused on both rural and urban community development through collaboration with local shuras. The humanitarian and development agencies viewed these direct links to communities as a way of bypassing a state that they saw as either ineffectual or illegitimate. This tendency grew even stronger under the Taliban. As it has elsewhere, the attempt to use aid as a substitute for politics and community development as a substitute for state building did not succeed. It did, however, strengthen the capacity of some communities to design and implement projects through some sort of participation. This capacity of community self-reliance, which one observes in other situations of state collapse, such as the Democratic Republic of Congo, can make eventual devolution of power more feasible. Such an outcome is probably preferable to a recentralization that once again deprives communities of participation in the country’s development.
When these regional power centers emerged, many observers, and some Afghans, saw them as the development of more grassroots forms of political organization that might provide a more effective and accountable alternative to the overly centralized Afghan state. During the period of the ISA, however, when the power of these regional coalitions was at its height, they became identified instead with warlordism, predation, and corruption. Faced with an “infidel” central government supported by the Soviet Union, Afghans had turned to local, tribal, ethnic, and regional alliances for resistance and had attacked the symbols of the central state with weapons received from the United States, Pakistan, Saudi Arabia, Iran, and China. When the mujahidin and militia leaders became identified with the destruction of the capital and predatory rule, sentiment began to shift, it appears, back toward a more centralized form of government. “Strong central government” became the Afghan term for “rule of law,” and this sentiment underpinned some of the initial support that the Taliban received.
The brief interim of Taliban rule, under the Islamic Emirate of Afghanistan, represented an attempt at recentralizing the state. Under the ISA the relatively impotent core of state power was held by SCN, that is, in most Afghans’ views, by the qawm of Panjshiris. But this qawm never consolidated its hold over the country, as did the Muhammadzais and the dynasties of ‘Abd al-Rahman Khan and Nadir Shah. The core of Taliban power was a group of Kandahari mullahs educated in Deobandi madrasas, most of them from rather low-ranking tribal backgrounds. Their link to the network of mullahs enabled them to expand around the country. The weakening by the war of competing elites facilitated their expansion.
Their social linkages, ability to deploy the legitimating discourse of Islam and to use sharia as a tool of rulership, together with the assistance they received from multiple sources in Pakistan (not solely the military) and eventually their Arab allies enabled the Taliban to recentralize power. They expelled the mujahidin and ex-communist warlords, abolished zones, and reconquered the country. Their ability actually to suppress opium cultivation illustrates the reach of their administration. Though they sailed through the Pashtun areas of the country with little fighting, and squeezed Massoud out of Kabul in a pincer movement, their conquest of northern and central Afghanistan was brutal, reprising the massacres of Hazaras carried out by Amir Abd al-Rahman Khan. They reestablished the practice of appointing governors and wuluswals from outside the areas they governed and were able to transfer them. The Taliban army broke up the smaller units based on solidarity groups led by commanders in favor of a more hierarchical system that, unlike any mujahidin or militia units, it could deploy anywhere in the country to concentrate forces against a threat. It also employed some turncoat qawmi units, but these were many fewer than among their opponents.
Initially, the Taliban’s expulsion of warlords and imposition of some kind of law, even if a brutal and simple one, won them support, at least in the Pashtun areas. The security of travel that they provided helped reinvigorate commerce, which mostly consisted of transit trade, or smuggling. They collected the customs revenues and remitted them to central control, though they did not seem to practice a very systematic form of record keeping, except where the experienced bureaucrats continued doing what they had done under every regime. In an interview in Kandahar in June 1998, Wakil Ahmad Mutawakkil, who later became the foreign minister of the Islamic Emirate and is now a prisoner of the United States, reportedly near Kandahar, stated, much like Muhammad Eshaq, that federalism as practiced, for instance, in Pakistan was unsuited to Afghanistan, which required a strong, central government. [In 2013 Mutawakkil was living in Kabul as one of a small group of prominent “reconciled” Taliban.] The ideal system he outlined, a highly centralized administration based on Islamic law, resembled a great deal the goal of Amir Abd al-Rahman Khan.
The U.S. and coalition offensive destroyed the Taliban’s centralized system. The United States made its first military alliance in Afghanistan with Fahim, who succeeded Massoud as military commander of Shura-yi Nazar after the latter’s assassination. The United States also supported other former regional commanders whom Massoud had brought back, in particular Dostum and Ismail Khan. In the south and east, where no single leader stood out, the United States supported some of the same commanders as in the past, including Hajji Abdul Qadir, whose brother, Abdul Haq, was executed by the Taliban and whose family had led the Jalalabad shura during 1992–1996; and Gul Agha Shirzai, who had been governor of Kandahar for part of that period and whose father, Hajji Abdul Latif, had been a commander of the National Islamic Front of Afghanistan and the Barakzai tribe in Kandahar.
The resurrection of the same regional commanders and warlords defeated and expelled by the Taliban created a situation in the provinces similar to 1992–1996. The major difference was at the center. Massoud had decided not simply to treat the state as war booty if he ever retook it, but instead to seek broader alliances, in particular with the former king, Zahir Shah, and through him with Pashtuns. This strategy bore fruit to some extent in the Bonn agreement, which legitimated a government led by a Kandahari Pashtun associated with the former king, Hamid Karzai, with the security organs under Panjshiri control, and other factions given other ministries. As mentioned previously, the legal framework for this government was the 1964 constitution minus the monarchy, abrogating the Islamist enactments of both the Taliban and the Islamic State of Afghanistan.
The actual distribution of power resulted largely from U.S. decisions about how to conduct the war, mainly by funding allied ground forces rather than putting U.S. troops in harm’s way. At the end of the initial campaign, U.S. Secretary of Defense Donald Rumsfeld reversed a previous decision by ordering Shura-yi Nazar into Kabul city. Previously, SCN had promised not to enter the city, and the Bonn agreement, signed after the city had fallen, requested an International Security Assistance Force to provide security there. The agreement provided for the withdrawal of all other military forces from Kabul, in order to remove the capital from factional control, but Fahim refused, and everyone else acquiesced. ISAF’s presence has done a great deal to enhance the national and neutral status of the capital, but the constant presence of Panjshiri soldiers and of pictures of Massoud reinforces a message about who is in control. Massoud may have decided not to treat the state and the capital as war booty, but this message did not reach some of his followers.
The result today is a mix of the legally mandated centralization with a military pseudo-federalism based on armed groups built up over years of war and reinvigorated by the United States and its coalition partners as part of the war on terrorism. The leadership in Kabul wants to recentralize the government but is split between advocates of a direct assault on regional forces and proponents of gradualism. The technocrats returned from the West tend to be in the former camp and the former Northern Alliance commanders in the latter. A few leaders, mainly outside the central government, propose federalism or decentralization. The international community involved in assisting Afghanistan in such matters pays lip service to the agenda of strengthening the central government but works pragmatically with a variety of regional forces, reinforcing fragmentation. Global trends that see decentralization as conducive to democracy and grassroots development legitimate such decisions.
Most Afghans seem to identify warlordism (jangsalari, tufangsalari) as the source of the country’s current problems. Ironically, some of the individuals whom others see as warlords advocate eradication of warlordism. Many people also agree on the creation of a national army and the emergence of a strong central government. Most people qualify their statements in favor of a strong central government and national army by adding that the current government in Kabul, seen as dominated by Shura-yi Nazar, does not enjoy sufficient national legitimacy to integrate the various regional forces.
The biggest change appears to be that most Afghans no longer accept the notion that the state is booty to be won by the strongest, provided that the ruling group deals justly with the people by enforcing sharia. Many Afghans today want to participate in their government as citizens of Afghanistan, even if they have widely different interpretations about what that means. If they reject control by unaccountable regional warlords in favor of what they call “strong, central government,” it does not mean they would accept an unaccountable, strong, central government, even if the latter somehow eradicated the warlords. Given the geographic and ethnographic realities of Afghanistan, such participation will require significant devolution of power, though not to the current regional armed groups.
Most regions in Afghanistan today are under the predominant influence of figures who exercise extralegal power, such as governors-general, combining military and administrative functions over several provinces. They effectively replace the central government’s authority, though they may have sworn allegiance to it. Whether they are officially provincial governors (Ismail Khan, Gul Agha Shirzai, Hajji Din Muhammad), military generals (Abdul Rashid Dostum, Ata Muhammad, Hazrat Ali), or incumbents of other offices (Vice-President Abdul Karim Khalili), they have gained their positions not by appointment from the center but because they command armed militia forces aided by the United States, in most cases affiliated with one or another faction of the former United Front (Northern Alliance). The remaining governors are also mostly former commanders from the provinces where they now exercise authority. Though all governors and generals have received letters of appointment from the president as required, they received these appointments because they have power, rather than the reverse.
The big so-called warlords have brought smaller armed groups of their province or region under their command or influence. Some poorer or weaker provinces acknowledge the suzerainty of a neighboring governor or regional commander, creating de facto zones. Ismail Khan in Herat and Gul Agha Shirzai in Kandahar are governors without official military positions, but they are simultaneously the de facto chiefs of the military forces not only of their own but also of their neighboring provinces. Abdul Rashid Dostum and Ata Muhammad are military generals without official administrative positions, but they have divvied up, and occasionally fight over, the civil administration in the northern provinces. They have appointed governors and other high officials and set policy in the provinces under their influence. A more complex case is that of the eastern region, where Hajji Din Muhammad is the governor of Nangarhar, and Hazrat Ali is the military commander of the Jalalabad garrison. Theirs is an uneasy condominium, as they belong to different factions and ethnic groups. The relative balance of power between the two may be due to the fact that the governor of Nangarhar enjoys local tribal support and some influence in neighboring provinces, while the military commander has strong backing from the defense minister in Kabul.
Further divergence from the old system is visible in the intraprovincial appointments and chain of command. Most provincial strongmen—be they governor or military ruler—appoint wuluswals. The governor of Herat, for example, has appointed all wuluswals of his province. In districts close to the provincial capital or with larger Tajik population, the wuluswals have managed to exercise control. In some districts, where the inhabitants are predominantly Pashtun, the governor’s Tajik appointees are accused of practicing gross ethnic discrimination. In Chisht, for example, there exists a power structure parallel to that of the official district government. In the district of Shindand, where Ismail Khan also attempted to appoint a wuluswal, the situation has escalated into open military confrontations between the governor’s forces and those of the local Pashtuns, under the command of Amanullah Khan. The latter has received some support from Gul Agha Shirzai, governor of Kandahar.
The governor of Nangarhar also claims that in most cases he has appointed wuluswals and district security chiefs who are not natives of their districts of assignment. Kandahar is a somewhat atypical case in that the governor at first experimented with district-level elections but interfered later. First, he authorized shuras in each wuluswali to elect the wuluswals. Three months later, the governor shifted all the elected wuluswals from their original districts to other ones.
The district of Spin Boldak, which is inhabited predominantly by the Achakzai and the Nurzai tribes, seems to have presented a particular problem. According to Kandahar high officials, the two tribes “could not resolve their problems and nominate a wuluswal.” The deputy governor explained, “Finally, the Governor sent a sayyid as their wuluswal.” The sayyid, a member of a religious descent group outside the Pashtun tribal structure, is originally from the district of Panjwa’i and was brought to Spin Boldak with the help of the governor’s armed forces.
The tribal chiefs in Spin Boldak, however, vehemently deny such a disagreement during the elections process and claim that the governor manipulated the process to serve his economic interests. According to a prominent Nurzai khan, “The governor does not have any particular tribal or ideological competition with us. The entire issue is over the control of commercial trucks.” In Spin Boldak, the governor has directly appointed all government officials from among individuals who do not belong to any of the two strong tribes of the region and come from other districts or even provinces.
On occasion, strong commanders have even appointed governors in the provinces in their zones of influence. In the immediate aftermath of the fall of the Taliban, General Dostum appointed the governors of Jauzjan and Faryab, who in turn appointed the wuluswals. Soon after taking office in February 2003, Interior Minister Ali Ahmad Jalali named new governors to those posts after consulting with Dostum, who reportedly raised no objection. When the new governor of Jauzjan tried to take up his post, however, factional fighting broke out. In Ghor province, where the governor was named by the central government, the deputy governor was appointed by Ismail Khan. The governor, who is from a district in the south of the province, could not even enter the provincial capital, Chaghcharan, which is located in the district to which his Herat-appointed deputy belongs. The deputy governor therefore conducted the affairs of the province with financial support from Herat.
This power structure has affected all elements of the state. The military units stationed in garrisons around the country are entirely recruited from those regions and serve the regional commanders. The Kabul garrison is almost entirely composed of Tajiks from Panjshir and Shamali. In this sense there is a factional rather than national army. At times the different military units fight each other, especially in the north.
The judiciary formally reports, as previously, to the Supreme Court in Kabul. That court itself is currently operating contrary to the law in several ways (size of the bench, age of the judges, rules of standing for considering cases). The provincial judges however, are subject to pressures and threats from the local commanders so that rendering justice is precluded in any case in which the regional commander has an interest.
The change in the operation of the revenue system is, of course, fundamental. The bureaucracy set up for tax collection by central governments in the past still exists, if in a weakened form, but it now generally functions not as part of a national state but under the control of regional commanders. Customs revenues have become a point of contention between Kabul and the provinces where major customs posts are located, as well as among rival commanders and provinces.
Since the mid-1990s Pakistan has banned twenty-some commercial items from passing through its borders into Afghanistan under the Afghan Transit Trade Agreement (ATTA), which provides for duty-free import into Pakistan of goods under seal for reexport to Afghanistan, which facilitates smuggling into Pakistan. After the ban, trucks carrying those contraband goods found new transport routes through Iran and Turkmenistan, into Afghanistan through Herat. This shift in transport routes has decreased the customs revenues of Kandahar and Nangarhar, while it has placed Herat in a markedly advantageous position. According to a high official of the customs service interviewed in Jalalabad, 70 percent of Afghanistan’s foreign trade was transiting through Iran in mid-2002. This will only increase with the new transit agreements negotiated with Iran by Afghan Minister of Commerce Sayyid Mustafa Kazemi in 2003. [Kazemi was assassinated by a suicide bomber in November 2007.]
Under so-called normal circumstances (which have not existed for decades), customs duties collected by the provincial customs office are deposited in the finance ministry’s bank account at the provincial branch of Da Afghanistan Bank. Then, according to the expenditure allocation for the province determined by the various ministries in their budgets, the ministry of finance in Kabul authorizes the provincial mustufiat (finance office) to withdraw money from that account and disburse it to the provincial governmental offices. If there is a surplus, the excess is transferred to Kabul. In case of shortage, additional funds are transferred from Kabul to meet the province’s budgetary requirements. According to Minister of Finance Ashraf Ghani Ahmadzai, Herat is the sole province that consistently enjoys a revenue surplus; all others rely on the central government’s revenues to cover their shortfalls.
In the past, too, customs revenue was rarely enough to cover all the expenses of a province. As the mustufis (finance officers) of Herat and Kandahar, emphasized, “Taxes collected at customs were never physically sent to Kabul.” Therefore, officials of these two provinces believe, they are not violating existing laws and procedures by not sending the customs revenue to Kabul. In reality, the governors of these provinces commit a gross procedural violation when they assume the role of the finance ministry in deciding on the disbursement of these funds.
There are similar difficulties in northern Afghanistan, where the major commanders split the revenues from the Hairatan customs house. Though the shares have been a repeated subject of fighting and renegotiation, by early 2003 the division of spoils, according to UN officials, seems to have stabilized at 50 percent for General Atta Muhammad of Jamiat, 37 percent for Dostum, 12 percent for Muhaqqiq and Hizb-i Wahdat, and the remaining 1 percent perhaps used for the cost of transactions.
This system has aggravated conflict among regions, particularly between Herat and Kandahar. “Ismail Khan is collecting taxes that are our rights,” said the deputy governor of Kandahar. The pre-1978 law says that a customs office can levy tax only on companies that are registered in that province. “Herat illegally taxes the companies that are registered elsewhere,” complained the Kandahar mustufi. Consequently, when a company registered in Kandahar brings its goods in through Herat, it pays duties in Herat. When its trucks reach Kandahar, that province’s customs office levies an additional tax on them because this is the province where the company is registered. If the final destination is Kabul, the company may pay import tax for the third time. Traders have staged protests to call attention to this problem on a number of occasions.
Ismail Khan, General Dostum, and, previously, the leaders of Hizb-i Wahdat (Khalili and Muhaqqiq) have supported a federal system of government. Partly because of the unprecedented degree of participation in the central government by Shia in both the interim and transitional administrations, and partly because of their recognition of continued resistance to federalism by Pashtuns, Hizb-i Wahdat has moderated its program to calling for measures of devolution and nondiscrimination against Shia jurisprudence. Over dinner at his residence, Muhaqqiq joked about Hizb-i Wahdat’s advocacy of federalism: “Dostum is with us with his four families [chahar khana], the Tajiks are reluctant, and the Pashtuns consider that federalism is kufr [unbelief] and that it is beyond the pale of the Islamic religion.”
Among the specific issues at stake in the debate over “federalism” are the zonal system (whether provinces should be formally grouped into zones or regions for administrative, fiscal, and political purposes); whether the leadership of the administration of provinces or zones should be elected, appointed, or some combination of the two; and what powers these subnational units should have, especially in the fiscal area, raising taxes and determining their own expenditures.
More generally, the issue is whether the current diffusion of power among various centers constitutes a positive devolution that can make the Afghan state more responsive, if these power centers are integrated into a legally functioning system, or whether they constitute inherently disruptive warlordism that should be suppressed by a strong central government.
Among political elites, those exercising power in the provinces and representing regions and ethnic groups that have long felt themselves to be subjects of a Pashtun-dominated state in Kabul rather than equal citizens of Afghanistan tend to advocate federalism or decentralization. Those exercising power in Kabul and Pashtun elites who see Afghanistan primarily as a Pashtun-led state advocate a strong, centralized government.
Popular perceptions seem to have evolved with changing experience. At the beginning of jihad, in 1979–80, before most fighters were organized into parties or organizations, they often spontaneously attacked the physical presence of the central government. People affiliated themselves with leaders largely from their own group who could provide security or dispense patronage. Now, however, after years of “gunlordism” (tufangsalari) by local commanders, popular sentiment has largely shifted in favor of a “strong central government.” The proceedings of the loya jirga in June 2002, which appeared to reflect popular sentiment quite closely, were notable for the absence of calls for decentralization or regional autonomy. The most popular theme was the need for the central government to subordinate or eliminate the “warlords” and provide security. One should not necessarily conclude therefore that government should be as centralized as in the past: confronted with an ineffective, corrupt, or oppressive central government, a possibility that one can hardly exclude, popular sentiment could shift once again.
President Karzai strongly opposes the zonal system and has insisted, for instance, that the consultations on the new constitution be organized in thirty-two provincial offices rather than in eight zonal offices, as the Emergency Loya Jirga elections were. Although he has compromised with the de facto decentralized power structure by confirming appointments of local officials from their own areas, in private discussions as early as Nawruz 1381 (2002) he confided his intention to return to the tradition of appointing governors and wuluswals from outside the areas they administer and said that this corresponded to popular demands.
Since his appointment as minister of finance at the 2002 Emergency Loya Jirga, Ashraf Ghani Ahmadzai, former senior social scientist of the World Bank, has emerged as one of the most forceful advocates of centralization. Since assuming office he has dedicated himself to trying to centralize finances and reform the customs service and treasury department of the ministry of finance. He has persuaded Ismail Khan to turn over new afs 100 million on several occasions and has held up payments to northern Afghanistan in order to pressure Dostum and the other northern commanders to turn the customs revenue over to central control. He has also held up payments to the defense ministry, insisting on a plan for reducing the size of the forces and clear identification of those whose salaries are being paid.
These demands have placed him in sharp conflict with Marshall Qasim Fahim, vice-president and minister of defense. Fahim does not challenge Ghani’s agenda of centralization. In an interview he stated: “A strong central government is imperative. The central government must have total control over the entire country; otherwise, it can’t be called a state.” What Fahim challenges is whether the centralization process should be used gradually to incorporate the armed forces of the Northern Alliance or to marginalize them in favor or a state structure dominated by a technocratic, and largely Pashtun, elite.
The governors and governors-general in the Pashtun areas also state they favor centralization, even if not under the current power structure. Gul Agha Shirzai of Kandahar states that he supports the establishment of a strong central government as the only means to a return to the rule of law, stability, and equitable economic development. Similarly, Hajji Din Muhammad, governor of Nangarhar, said: “Federalism is not the right formula for Afghanistan. The federal systems that are in place in Pakistan or in Germany will not be successful in Afghanistan. They would only push the country towards disintegration.”
The argument that federalism will lead to disintegration derives from the experience of the past decades, in which regionally based commanders developed independent patronage and supply relations with governments and other support networks in neighboring countries. These political and economic linkages provided the infrastructure for the proxy war that took place on Afghan territory, pulling the country apart. Of course, in genuine federalism, armed force and security would still be centralized.
Sometimes Afghan commanders and governors-general speak as if federalism were a system where each zone would have its own army. This is not federalism, but the association of the term with these facts on the ground colors how Afghans perceive it.
The differences over state structure are not only about the goals, but about the process. The conflict over the nature of the process is most intense between Ashraf Ghani and Marshall Fahim, but it expresses the difference in viewpoint between the returning technocrats and those associated with previous regimes, on the one hand, and those whose road to power was through the jihad and the struggle against the Taliban.
The former position advocates a kind of “shock therapy” centralization. Its advocates seek to use international financial assistance to the central government as leverage to subordinate the regional or zonal authorities, construct a centralized administration, and win the loyalty of the people and former fighters to the government. A Kabul-based elite with a heavy admixture of Western-trained and Western-returned technocrats would exercise most political power.
The gradualist group largely associated with the former Northern Alliance, however, seeks to incorporate the de facto decentralized military and administrative structures, headed by members of their own and allied organizations into a more centralized structure. An elite composed of former commanders and regional leaders (warlords) would share rulership with Western-trained and returned technocrats while retaining a predominant, though not exclusive, role in the security forces.
Two cabinet ministers in Kabul who participated in the Bonn Talks as members (in one case as leader) of different delegations offered intermediate solutions that accepted the need for a process of state building that would transform current realities gradually and keep evolving.
Muhammad Yunus Qanuni, minister of education and former minister of the interior, who led the United Front delegation in Rome, expresses the gradualist point of view: “The debate should not be on either an all-powerful central state or a federal system. Neither is realistic at the moment.” Qanuni suggested that a series of authorities should go to the center, while the remaining one should be given to governors.
For example, all level I and level II [rutba-yi awal wa rutba-yi du, the highest ranking civil servants] officials should be appointed by the center—including governors and wuluswals, and the rest by the governors. Fifty percent of customs revenue should be sent to the center. Teachers should be appointed locally. All this, until a strong central state is established.
M. Amin Farhang, minister of reconstruction, who was a member of the Rome delegation, argued:
A federal system cannot be established in the short-term. Currently escape from the center is a reality, but this will take Afghanistan towards gradual disintegration. The muqam-i sarkari [state authority] must be reinstated. Only after a strong central government is established can we slowly begin to move towards decentralization.
In this view, a central power must be established, before it can be devolved. The illegitimate regional power centers must be dissolved or subordinated, before laws on new forms of devolution can take effect.
For now, centralization is popular as slogan or symbol among the common people throughout the country, but it is far from certain that the popularity of this idea would long survive its implementation. This view reflects people’s strong distrust of the existing local and regional rulers, whose use of ethnic, regional, and tribal grievances to legitimate their power people seem largely to reject. Even after a quarter of a century of war and disintegration of the central government, state structures and procedures that were deeply engraved in the minds of the people and of the cadre of remaining bureaucrats were quickly revived in both the center and the provinces and districts. New leaders have changed the rules of politics, but administrative structures and bureaucratic procedures have stubbornly remained the same.
The popular views of government, however, may not be as simple as they appear. Decades of conflict, exile, and politicization have created an Afghan public quite different from that of several decades ago. Many Afghans want to be participating citizens, not passive subjects. One author (Rubin) asked an ad hoc assembly in a mosque in a village near Bamiyan how the relations of the center to the provinces and districts should be structured. One local man who was clearly a leader in the community responded that the people should elect the wuluswals, and that the central government should appoint the governor from a list of three candidates submitted by the local people, presumably by a provincial assembly. We are confident that others have such ideas as well, which would help the Afghan political debate move beyond the sterile polarization of “federalism” versus a “strong, central government.”
The drafting of the new Afghan constitution should provide an occasion to debate these issues more fully, though merely adopting a plan for the structure of the government will not guarantee its implementation. Probably the best constitutional framework would be a unitary state structured in such a way as to provide for participation at all levels through elected councils and that is flexible enough to permit extensive devolution of power to provinces, districts, and localities. The desirable amount of devolution will vary with the sector and over time.
Popular sentiment clearly wants a more consolidated state right now, and particularly the demobilization of militias. To the extent that Afghanistan has an army, it must, of course, be centralized. Policing is an area ripe for new models, at least once a minimally professional national gendarmerie is reestablished. Tax collection will have to be largely nationalized, but provinces could be allowed certain revenue powers. More important, given the great disparity in resources among provinces, the government should consider establishing an institutional structure that will make it possible to give block grants to provinces, districts, and communities to budget as they see fit for certain purposes. Elected councils could be given jurisdiction over such allocations. The idea of appointing governors and wuluswals from the center is deeply ingrained but can at some point be reconsidered. Some form of zonal system may be necessary for economic and social planning, if not for political participation. It would be remarkably inefficient if every national program had to be administered through one office in Kabul and thirty-two provincial offices, rather than through eight or so regional offices.
All these topics can be the subject of rational debate, but the political issues are those that might plunge the country back into conflict. A weak and largely unaccountable state dependent on foreign aid and foreign advisors before 1978 nurtured a political elite that was concentrated in a few neighborhoods in Kabul and increasingly culturally and politically distant from the rest of the country, which it barely controlled through a highly centralized but ineffective administration. Today’s rapid recentralizers and their foreign backers risk creating a similar situation. They also risk reigniting conflict if they too aggressively attack well-entrenched regional powers, before they have alternatives ready for the fighters and, even more important, their leaders. Already one hears that little aid leaves Kabul and is mainly used for international organizations, foreigners, and a small clique of Westernized Afghans who speak their language. Such a system will be even more unsustainable in the Afghanistan of the twenty-first century than in the previous one. Today’s warlords could be tomorrow’s champions of revolt against the foreign imposition of a small, unaccountable elite in Kabul. Afghans do not support the regional gunlords, but they will not accept rule by aid-lords either. The political task of reconstruction will be to help Afghans use the assistance they need to construct a state structure that is unitary enough to unite the country and decentralized enough to permit real participation to a population that is politicized as never before.