On 22 December 2001, following the defeat of the Taliban by a coalition of US and Northern Alliance forces, an interim administration under Hamid Karzai mandated from above by the Bonn peace agreement took charge of Afghanistan. Being a relatively small, landlocked and poor country that has experienced civil strife for more than two decades starting 1979, with different degrees of external involvement, Afghanistan's economy at the time of formation of the interim administration was both backward and disintegrated. With 85 per cent of its population reportedly dependent on agriculture and 53 per cent of the GDP estimated as originating in the agricultural, livestock and forestry sectors, as compared with 28 per cent in light industry, 8 per cent in trade and 6 per cent in construction, the economy in 2002 was structurally in an early stage of development. Extremely rough estimates quoted by the Asian Development Bank suggest that in 2002, Afghanistan's GDP amounted to $4.4 billion, which implied a GDP per capita of an abysmally low $170 per head.

These figures, however, could exaggerate the level of backwardness implicit in the initial conditions from which Afghanistan must begin its process of reconstruction and revival. Two factors have contributed to the high 'aggregate poverty' indicated by a per capita income that is less than half of the international poverty norm of a dollar per day per head by 2001. The first is the war that led to the ouster of the Taliban. An earlier estimate, also quoted by the ADB, relating to 1989, placed Afghanistan's GDP at a much higher $6.9 billion and its per capita income at around $300. The situation could have further improved in the years following 1989, since reports indicate that at least in regions fully occupied by the Taliban, economic conditions were stable during the early and the mid-1990s. Underlying the subsequent massive contraction of the economy was the war that devastated the limited infrastructure of the country, triggered the exodus of more than 3 million refugees to Pakistan, Iran and elsewhere and displaced a large number of people within the country. This disrupted or even brought to a standstill much of the economic activity within the country. A corollary of this role of the late 1990s war in worsening economic condition is that a concerted reconstruction effort focused on quick-impact projects combined with the observed large-scale return of refugees to Afghanistan, could ensure a sharp rise in GDP and per capita income to levels in the early 1990s.

Secondly, nature too added to the woes of an already war-ravaged economy, with protracted drought conditions that began in 1999 and lasted till 2001. Again according to estimates made by the international community, in this case the World Bank, crop production was halved and livestock herds substantially depleted, making the situation in 2002 as desperate as it was. According to a March 2002 survey by the FAO, the livestock population in the country declined by 60 per cent due to the distress selling of livestock herds in the summer and autumn of 2001 triggered by the persisting conditions of drought. The crucial role of drought in worsening economic conditions is revealed by an FAO and World Food Programme estimate which states that agricultural production in 2002 rose by 82 per cent, following better rainfall conditions. However, even in that year cereal production is estimated at 4 per cent below its 1998 level. Given the importance of agriculture for employment and GDP in the economy, these figures suggest that the rough GDP estimates relating to 2002 may be exaggerating the poor state of the economy.

These real factors that perhaps exaggerate the extent of backwardness of the economy could have been compounded by statistical errors that remain uncorrected because of the temporary closure of the statistical system. It is known that GDP and revenue estimates in Afghanistan are quite unreliable given the large role of the unaccounted, underground economy in the country. Two factors contribute to the magnitude of the unaccounted segments of the economy. To start with, Afghanistan, being a landlocked country, is also a hub for trade between some of its larger neighbours. Such trade flows occur through its porous borders with Pakistan (2,450 km), Tajikistan (1,206 km), Iran (936 km), Turkmenistan (744 km), Uzbekistan (137 km) and China (76 km). Unfortunately, figures on Afghanistan's domestically directed and re-export trade are not available. A June 2000–February 2001 UNDP study, based on an investigative survey of border crossings and major trading centres, estimated that 'indigenous exports' at $130 million constituted just 10.6 per cent of total exports, while re-exports totalled $1097 million (Chart 1). Similarly, imports for domestic consumption, based on type of commodity, amounted to $396 million (33 per cent), whereas imports that were potentially for re-export totalled $806 million (Chart 2). It is to be expected that the huge volume of transit or re-export trade would sustain an economy whose activities are not fully captured in estimates of GDP tracking the real economy. The fact that much of this trade is not officially accounted for is well known, so that profits accruing to Afghan nationals from such trade (in which exports account for more than a quarter of estimated 2002 GDP) are also likely to be inadequately captured by national income figures.

The second reason why the magnitude of the unaccounted economy is likely to be great is the historically important role of poppy cultivation and opium production in the rural Afghan economy. According to an estimate by the United Nations Office for Drug Control and Crime Prevention, land under poppy cultivation amounted to 82,171 hectares in 2000 and opium production in 2002 was 3,400 tons, which was similar to levels that prevailed in the mid-to-late 1990s. With the Taliban having imposed a ban on opium production and the Karzai government having declared a ban on poppy cultivation in 2002, it is inevitable that a large part of this activity would occur in the underground economy. The economy surrounding the cultivation of poppy, the production of opium and its trade would generate a significant income, which again will go substantially unrecorded.

Ignoring the potential areas for rapid reconstruction of war-damaged assets, the effect of prolonged drought and the important role of unrecorded economy tends to exaggerate the poor initial conditions, which are then attributed solely to the long years of conflict. This is not to say that the two-decade-long war had no serious consequences. To start with, it did have damaging effects on the infrastructure of the country, created by the historically important role of foreign aid in the country because of its strategic and transit trade importance, making agriculture more climate dependent, fragmenting the economy by damaging the transportation and communications infrastructure and generating bottlenecks in sectors like energy. Secondly, by driving out large sections of the population and emasculating the state it undermined the institutions of state including the central bank, the financial system, tax collection and customs machinery, the statistical apparatus, the civil service, and the law and order and judicial system. Finally, it undermined the ability of the state to collect a share of the surplus being generated within the economy to finance the expenditures necessary to take on crucial governance and development responsibilities.

With regard to the last of these effects, we must recall that local revenue in Afghanistan comes from its taxes on trade, much of which occurs by land across the borders with its neighbours. A major problem being faced by the current government in Kabul is to get provincial governors (still referred to as warlords by many) to part with a reasonable share of the revenue thus garnered. Even though a meeting in May saw twelve governors signing an agreement to do so, and one of them delivered $20 million immediately thereafter, it is unclear how much of the revenue will finally accrue to the central government.

Once these features of the Afghan economy are understood and the unwarranted hopelessness deriving from perceptions of near-complete devastation and penury is set aside, the direction of the aid-supported reconstruction effort and its likely consequences can be easily deduced. There are three pillars on which that effort should rest: (i) restoration of the infrastructure with initial focus on quick-impact projects that can yield substantial benefits in a short period; (ii) restoration of the institutions of state and the powers of those institutions so that the state can mobilize the revenues and undertake crucial developmental expenditures, which only it can pursue since low per capita income levels imply inadequate incentives for private investment in many areas; and (iii) restoration of the 'economic space' within which the new state can pursue a national development agenda.

A range of specific initiatives can be identified within this overall framework. To begin with, a greater degree of security and an end to periodic local conflict must be ensured. This must build on the intense and easily observed desire of the Afghan people for peace and the opportunity to get on with their lives and must be secured only by an international, UN-mandated force that is seen as a source of support and not an instrument of occupation. Second, the damaged and destroyed infrastructure must be immediately reconstructed, with a large proportion of aid diverted for the purpose, so that the economy is reintegrated and rendered functional.

Third, the ability of the state to mobilize resources needed to finance crucial developmental responsibilities must be restored for which an appropriate monitoring, tax collection and revenue sharing system must be worked out. Given the fact that Iran and Pakistan account for a substantial share of Afghanistan's trade, and trade revenues constitute a large share of the total, the initial task should not be too difficult. But over time, it would be necessary to widen the tax base and obtain resources from internal direct and indirect taxation, especially the former.

Fourth, it would be necessary to rebuild the financial system with a two-fold purpose: that of mobilizing household savings and of channelizing those savings to priority areas, using mechanisms such as directed credit and differential interest rates. This implies that an overemphasis on microfinance and any neglect of the task of creating a well-regulated formal banking system would be misplaced. Fifth, the rural infrastructure needed to restore dynamism to agriculture, reduce dependence on rainfall for irrigation and encourage growth of non-agricultural activities needs to be created by the state.

Finally, Afghanistan needs to move out of being a mere transit hub for trade between its neighbours and must seek to develop activities that add value to imported inputs to produce outputs for export and benefit to a greater extent from its crucial role in regional trade. The problem with the transit trade is not that it adversely affects the balance of payments. The field survey-based estimates quoted above show that though there was a small deficit in terms of the indigenously directed and originating trade, Afghanistan did not suffer from a trade imbalance that needed significant external financing. The real difficulty is that unregulated growth of such trade creates disincentives for investment aimed at increasing domestic employment and value-added through processing and production activities within Afghanistan.

Is this the direction in which the ostensibly UN-coordinated reconstruction effort is heading to? Expectations were high when, soon after the Bonn agreement that signalled peace, a meeting of donors in Tokyo in January 2000 pledged to deliver sums, estimated at an aggregate of $4.5 billion within a thirty-month period, to kick-start the reconstruction effort. But a visit to Kabul, the capital city, close to a year-and-a-half thereafter suggests that the process of reconstruction has been slow. The inadequacy of the reconstruction effort is puzzling also because the limited and scattered information on aid flows does suggest that inflows have been significant even if not massive.

Aid flows take many forms. They occur directly to the government, through the World Bank administered Afghan Reconstruction Trust Fund, the Law and Order Trust Fund (LOTF, which funds the police), the Afghanistan Trust Fund (that pays for the army) and a number of other channels. They occur through routes indirectly linked to the government, inasmuch as local and international NGOs are supported by donors to undertake projects that are in keeping with the National Development Framework adopted by the government in consultation with the donor community. They also flow into areas completely outside the governmental framework inasmuch as US 'coalition forces', other than the International Security Assistance Force (ISAF), and international NGOs bring in their own funds and also in cases where it is provided by donors abroad for activities that are not necessarily reported in full.

The most easily accessed information is on money that is directly contributed to the government. To coordinate these flows, donors have constituted the ARTF in response to a government request that the budget be made the central instrument to identify and allocate funds for projects. From the government's point of view, a link with the budget would help it guide and take credit for the reconstruction that aid permitted. It would also ensure that aid could be used to meet the recurrent expenditure of the government, including a large part of its wage and salary bill. This request was, in turn, appreciated by the donor community, because it ensured 'national ownership' of the policies that open or covert aid conditionality prescribed and it ensured a degree of transparency about the allocation and use of aid flows.

Moreover, with the ARTF being administered (for a fee) by the World Bank, which sees in the Fund a means of appropriating the policy leverage that the combined aid of numerous donors provides, powerful groups within developed-country donors were sure that their interests would be protected. In principle, the mandate of the ARTF is extremely wide. Besides funding the government's recurrent expenditure including its salary bill (which though anathema in the Bank's perspective is a source of substantial leverage over a resource-starved state), the ARTF can finance investment activities and programmes, including quick-impact recovery projects, capacity-building projects and payments made to Afghan expert-residents abroad who are willing to temporarily or permanently relocate to their home country. The last two of these are also a major means of influencing the decision making of the government.

Indications are that the Bank-led donor effort to direct aid flows to the government and coordinate those flows through the agencies like the ARTF and the Afghan Assistance Coordination Authority (ACCA) has progressed substantially. During the first full financial year (FY) 2002–03 under the new administration (coinciding with solar year 1381 stretching from 21 March 2002 to 20 March 2003), which came soon after the Tokyo meet and before the aid coordination framework could be put in place, the budget was not the principal means of centralizing aid flows and utilization. In April 2002, the interim administration adopted a recurrent budget for FY 2002 involving expenditures of $483 million, including $23 million for clearance of wage arrears. With domestic revenues estimated at $83 million, aid was expected to finance $400 million or 83 per cent of the budget. However, figures collated by the ACCA indicate that total grant disbursements during the last quarter of SY 1380 and the full SY 1381 (FY 2002) amounted to $1.84 billion, of which $295.9 went to the government through the budget, the ARTF, the Law and Order Trust Fund and other channels. The channels of disbursement and the local target agency for the remaining money are unclear, but a substantial part ($700 million according to one estimate) reportedly went to finance humanitarian assistance in the context of the drought.

It is in the financial year starting 21 March 2003 (SY 1382) that the process of coordinating aid flows is being consolidated. In March 2003, the transitional government of Afghanistan announced a comprehensive budget for FY 2003—comprising an ordinary (recurrent) budget of $550 million, of which $200 million is expected to be financed with local revenues and $350 million by external assistance, and a development budget of about $1.7 billion. According to ACCA, by June, financing from donors of $211.2 million for the recurrent budget and $1.2 billion for the development budget had been confirmed. Interestingly, only the $211.2 million for the recurrent budget has been channelled through the ARTF, the LOTF and the Afghanistan trust Fund, whereas grant funding for development programmes in the National Budget has come through other channels. The process of using the World Bank to enforce conditionality appears to work through a mechanism in which the Bank controls funding for the crucial and politically sensitive recurrent budget and thereby is able to influence both the government's policies and its utilization of the remaining aid provided directly to the government.

Given the funding for FY 2003 confirmed so far, expectations are that the budgetary target for external financing is likely to be fulfilled. The evidence collated by the ACCA suggests that aid to the government through various channels would by the end of FY 2003 amount to $ 3.36 billion. Thus over a period spreading over twenty-six months after the Tokyo meeting in January 2002, grants disbursed to the government would cover 87 per cent of the pledges made for a thirty-month period after the meeting and 75 per cent of the total pledges made at and after the Tokyo meeting. According to ACCA sources, it is this high rate of disbursement that has made the Karzai administration bid for a $15–20 billion aid package over a five-year period.

Though information on the flows through channels not directed at the budget and monitored by the ACCA is as yet difficult to come by, the actual disbursement of aid is likely to have been substantially more than reported above. Even though a variety of sources in Kabul believe that the sums involved would not be as large as that provided to the budget, it is likely that such flows would significantly add to the total. Thus clearly, overall disbursements have been in keeping with the expectations generated at Tokyo.

What then accounts for the picture of a slow and clearly inadequate pace of reconstruction? There are two obvious reasons why aid flows have not worked to refurbish the infrastructure, even in Kabul. To start with, as noted earlier, a large part of the flow that occurred prior to FY 2003 went to support the humanitarian assistance programme to deal with the consequences of the protracted drought and took the form of food-aid, much of which neither involved any capital spending nor involved expenditures in Afghanistan, since the food was imported. Second, a large part of the grant disbursement even in FY 2003 has gone to support the recurrent expenditures of the government, especially its wage and salary bill, which while spurring domestic consumption spending, has not contributed to savings and investment.

This use of aid to support the recurrent expenditures of the state, including its wage bill, has a number of implications. The inability of the government to finance its own recurrent expenditures reflects the fact that government in Kabul does not as yet have the legitimacy or the power to garner a sufficient amount of revenue that is reportedly being collected at the provincial level. A very large proportion of the government's 240,000 non-military personnel are located in Kabul. This creates a situation in which the government in Kabul is supported by the donor community, whose leverage is therefore substantial. Simultaneously, the expatiate presence and spending in Kabul has generated a services-and-construction boom there, which is bound to widen differentials between Kabul and the rest of the country. In a country riven with ethnic and other divisions, this distancing of Kabul and its identification with a foreign presence could further undermine the legitimacy of the Kabul government, making it difficult for it to garner the revenues it needs to do away with dependence on aid. A consequence is that cycle of dependence of the government on aid would only be strengthened.

The second set of reasons why aid has not created the expected impact is related to allocation decisions. As Chart 3 indicates, budgetary allocations for SY 1382 (FY 2003) indicate that while social sectors like education, health and refugee rehabilitation are allocated a total of $775 million and another $210 million is to be spent on foreign investment promotion, technical assistance and capacity-building in the areas of trade and investment, public administration and economic management and justice, only $509 million are allocated to the crucial areas of transport and energy, and mining and telecommunications. When seen in terms of aid commitments to finance these expenditures as of June 2003, this bias appears even greater (Chart 4). Thus part of the reason for slow reconstruction of infrastructure is that a small part of even the development budget is being allocated for the purpose.

This overemphasis on technical assistance and capacity building is justified in the name of reconstituting the damaged Afghan state. Much of the money under these heads goes to institutions like the Adam Smith Institute and expatriate professionals who are paid huge consultancy fees in foreign currency, not all of which is even spent in Afghanistan. A constant refrain heard among the expatriate population in Kabul is that Afghans lack the capacity to manage their own affairs. Despite the fact that these institutions and most NGOs manage their programmes, especially those outside Kabul, with the aid of local partners, and though many skilled personnel who left as refugees to Pakistan, Iran and elsewhere have returned, this perspective overwhelmingly influences the management of aid-financed programmes. It almost appears that the "lack of capacity" discourse supports a nexus of aid-financed professionals, who find in such arguments the justification of their own dollar-funded presence in the country, and of aid donors, especially agencies such as the World Bank, who can use those arguments to prevent the entry of those more nationalistic in orientation into the decision-making and implementation process. This is not a fact that goes unnoticed among the Afghan population. Interviews with Afghan professionals and members of the faculty at the University of Kabul inevitably generate comments to the effect that Afghan talent, more attuned to the socio-cultural and economic context of the country, is being ignored, while expatriate and non-resident Afghan "experts" are imported even when not required.

The still-evolving policy framework in Afghanistan provides reasons why nationalistic perceptions may be considered dispensable. It is to be expected that with government expenditure substantially driven by aid disbursements, explicit and implicit aid conditionality is bound to influence the policy framework. The effort, therefore, is to ensure that those manning the government "own" the policies that such conditionality implies. Unfortunately that policy framework is not of a kind that could support the creation of a "domestic economic space" within which the Afghan state can fashion a nationally beneficial and egalitarian development strategy. Though the government's donor-driven strategy is yet in the making, it is clear that the inevitable import dependence of a war-devastated economy is being intensified, by encouraging an open trade policy that would discourage domestic industrial investment. Crucial social sector areas like health provision are being handed over to private entities and NGOs to run at a price, which though currently subsidized by aid, could easily be charged to the user at a later date.

The state has adopted as law the principle that the central bank cannot lend to the government. Thus deficit-financed expenditure to revive the economy is ruled out, resulting in a near-complete dependence on aid for budgetary finance. The battered publicly-owned financial sector, rather than being supported and strengthened, in order to be leveraged for domestic industrial financing, is soon to be subjected to competition from private, especially foreign, players. Since the need for a banking system that can mobilize local savings and channelize it to priority investments crucial to domestic private sector development is great, this shift can be disastrous. Foreign banks would only wean away the most lucrative (mainly expatriate) businesses from the local, publicly owned banking sector, while refusing to take any responsibility for development. Finally, though the new Afghani is formally the national currency, aid flows and the expatriate presence is dollarising the economy with the dollar being traded on the streets and many service establishments quoting dollar prices and expecting payment in dollars.

These developments proceed unchallenged partly because a vocal, urban domestic elite is being incorporated into the aid economy. Not only are salaries paid on time with the help of aid, but the demand for services from the expatriate population is on the rise. Rents quoted in dollars are rising fast because of the expatriate rush into Kabul and some other cities. These and other opportunities are resulting in the emergence of a new class of richer Afghans, living off the aid economy, which is happy with the freedom that an open, dollarised economy provides.

All this is fine at the moment since aid flows are leading to foreign exchange reserve accumulation with the central bank, which can use those reserves when required to prop up the new Afghani and prevent local prices of imported commodities from rising. But in the medium term, aid, import-dependence and dollarisation can all prove a burden. If the pace of aid inflow slows, not only will the just-reviving economy contract but the Afghani would depreciate pushing up the prices of essentials in that segment of the economy still earning and living off the local currency. Stagflation, in an already depressed economy, would be the result. On the other hand, if aid inflows continue at present levels or rise in volume, inequalities between Kabul and the rest of the country, across differentially endowed regions and between income classes are bound to increase. That can have dangerous implications in a country that is still scarred by a complex chain of civil strife influenced by ethnic and religious divides.

The current model of development in Afghanistan is clearly unsustainable for any discerning observer. But for those present in the country with short-term military and economic interests in mind, who are protected by the façade of altruism that a small dose of aid to a poor and devastated nation helps provide, this appears to be of no concern.

The author wishes to gratefully acknowledge support from the ActionAid Regional Bureau, Bangkok and the Action Aid Afghanistan Office for the research on which this article is based. He also benefited immensely from discussions with Bijay Kumar, Essa Shamal, Phillippa Sackett, S. Parasuraman and P.V. Unnikrishnan.

July 23, 2003.