War, money and aid
by Roger Persichino May 2003

That relief aid has an impact on war economies – and at worst is guilty of fuelling or prolonging conflict – has become almost an article of faith within the humanitarian community. Yet this has been accepted in the absence of any real monetary analysis. Yet a close look at the economics of money in relief yields some surprising results.

Over the last decade, a great deal of attention has focused on the impact of relief aid on local power structures and the political economy of conflicts. Although the impact of relief is traditionally hard to assess, most observers now believe that, in one way or another, assistance fuels conflict. In other instances, aid has been said to distort local prices or salary scales. Comparatively less attention has been devoted to analysing the actual economy (or economies) in question. By looking at specific examples of actual war economies, this article aims to show that the adverse impacts of aid may in fact be overstated.

The valuation of aid: the case of Afghanistan

One way to gauge the real impact of relief assistance in conflict areas is through the valuation of aid relative to other sources of money within the economy. Thus, in Afghanistan we can compare the local value of aid with the value of opium sales and local trade under the Taliban.

In 1998 and 1999, according to the UN Office for Drug Control and Crime Prevention, Afghanistan produced about 2,500 tonnes of opium, the proceeds from which were said to fund the Taliban regime. A kilo of opium exiting Afghanistan was sold at about $10. Thus, total opium production could be valued locally at about $25 million per year. If we assume that the Taliban received 50% of this amount in ‘tax’ – a high-end estimate – that would imply that the opium trade earned the regime $12.5m a year. In comparison, in 1996 and 1997 aid to Afghanistan was worth $250m a year, according to the UN. If we assume that, at a minimum, 5% of this amount was diverted by the Taliban, then that would give us a figure of, again, $12.5m. In this model, the Taliban’s revenues from taxing opium at 50% equal its revenues from the diversion of 5% of total aid, although it is certain that the actual revenues accruing to the Taliban were in fact lower than this.

To put these figures into perspective, according to the World Bank some $3 billion-worth of trade, excluding drugs and guns, was done each year in the Kabul Valley, roughly between the Kabul-Jalalabad area and the tribal zones in Pakistan. This is over ten times the amount of aid for the whole country, and over 100 times the local value of opium production. Again, the actual taxable amount on which the Taliban could gain revenue is impossible to estimate, but is certainly much lower than the overall total turnover. Nonetheless, since the figures for opium production and local trade both signify turnover, we can compare their relative magnitude. There is no doubt that, to an extent, aid financed the war economy in Afghanistan. However, the total aid that entered Afghanistan paled in comparison with the size of the local economy, and its impact was dwarfed by the local economy’s contribution.

A related point raised by the Afghan example is that, for any analysis of the potential impact of aid on a war economy to be meaningful, aid has to be valued locally. In Afghanistan, aid, and the World Food Programme’s food aid in particular, are valued at the point of origin, and include costs for shipping and logistics which have little relevance to local economics. The actual local value of aid is therefore substantially lower, further decreasing the potential contribution of relief to the war economy.

The velocity of money: the case of Somalia

The following example, concerning currency fluctuations in Somalia, shows that we also need to take into account the speed, or what economists call the velocity, at which money, including aid money, circulates within the war economy. In other words, it is not only a question of the value of aid relative to other sources of revenue, but also of how quickly and in what ways aid funds are circulated and used.

Currency fluctuations in Somalia have been erratic since 1993, though broadly speaking depreciations in the Somali Shilling against the US dollar can be traced to fresh currency inflows in the large markets of Mogadishu, via transactions by a handful of local businessmen. The total amount of money introduced into Somalia in this way in 2001 has been estimated at between $30m and $50m. It is uncertain whether this amount was directly related to the introduction of newly-minted currency or to the re-injection of other money (for instance laundered cash); probably a combination of the two. For practical purposes, it can be considered as an inflow of new money relative to the local market. In comparison, remittances reaching Somalia through the Al Barakaat money-transfer service until its closure in November 2001 were estimated at $250m to $300m yearly. Remarkably, although remittances were so much larger than fresh currency inflows, they accounted for virtually no devaluation in the Shilling against the dollar on local foreign-exchange markets.

This apparent paradox is explained by the differences in how the two types of inflow actually reach the market. The total amount of remittances is indicative of Al Barakaat’s financial turnover, but can also be interpreted as the aggregate sum of many small individual holdings. This sum is actually split over time between thousands of small households in several distinct geographical areas, all linked to numerous distinct local markets. In turn, these households can save part of this money, and spread their expenses over time. This results in a slow diffusion of each holding into the local markets. This steady inflow of currency has little effect on the Somali Shilling’s value. Conversely, currency inflows channelled by businessmen reach the money market in bulk. There is no gradual diffusion process, and monies are tied not to a multitude of small markets, but to one or a few large ones. This means that this inflow is much more disruptive, hence the observed depreciation of the Somali Shilling against the dollar.

Both types of currency inflow are foreign-exchange transactions. In that sense, they are of a similar nature; where they differ is in how they are actually used. Remittances are used as a means of payment (with savings held for future payments by households), while ‘businessmen inflows’ are used for financial transactions. This distinction suggests that the economic concept of velocity – ‘a statistical averaging of money that moves with money that is mostly at rest’ – may be a useful tool in analysing the impact of aid on a war economy. Crudely put, velocity measures the speed at which money circulates in an economy. The example above discussed how two monetary inputs entered the Somali markets at a different velocity, with very different effects on the local economy. In the Somali context, the velocity of remittances is much smaller than the velocity of fresh currency. Although much more difficult to quantify, it is likely that the Afghan Hawalas, operating under almost the same procedures as Al Barakaat, lend themselves to the same conclusions and analysis. Essentially, small amounts with high velocity have a substantially larger impact than large amounts with low velocity.

The velocity of aid: the case of southern Sudan

Either through distributions targeting individuals or households, or through the payment of salaries to national staff, relief disbursements are by nature discrete, and in that sense similar to the Al Barakaat remittances. In most cases, they can thus be expected to have neutral velocity relative to the local economy. In other words, they are not distorted by velocity-induced effects.

However, as the case of south Sudan shows, this is not always the case, nor does it account for the micro-effects of aid when distributions are concentrated in one geographical area. The economy, or rather the economies, of southern Sudan are very different from Afghanistan or Somalia. First, they rely on barter or other forms of valuation, such as cattle in Nuer and Dinka areas, while cash trade predominates in the areas bordering Uganda. Second, the geography is hostile and the transport network, where any exists, is at best dilapidated. This has the dual effect of inducing a very low velocity baseline and fragmenting the economy into several small units concentrated around major local markets. By contrast, aid has a much greater velocity. At the local level where distributions actually occur, it is extremely likely that aid has a greater impact on the local economy in terms of relative volumes than it would have in other contexts. This has the well-known effect of making any distribution site a potential target for raids.

In cash-based areas, relief items can be sold on the market or confiscated. The analysis developed above for Somalia and Afghanistan is thus likely to hold. However, this cannot apply in most of the northern regions of southern Sudan, where the local economies are not monetised. Local market constraints obviously prevent any meaningful transfers from one area to another to the benefit of a faction engaged in the conflict. Aid must therefore be consumed locally, which can only benefit a small proportion of any given warring party. If we assume that most aid is concentrated in only a few areas, it may follow that the impact of aid is uneven across a war economy, and so needs to be assessed in a more localised way. The concept of ‘economy zones’, introduced in relief work by Save the Children, might prove a useful starting point from which to define the relevant geographical units within which such measurements could be made.


Standard tools of monetary analysis can usefully address the question of whether aid delivered into a war economy has a quantifiable impact upon the finances available for, or accessible through, war. Specifically, valuation based on local rates can be used for comparative and policy-making purposes. A careful examination of the velocity of aid relative to the local environment would allow for considerable fine-tuning of the analysis. Standard tools for this purpose exist for peacetime economies, and could be applied to conflict economies at macro and micro levels.

A further issue is to do with timing. For example, it is likely that, in the crises in Kosovo or East Timor, there was a very high level of inputs in the first months of the response, and a lower level of inputs in later months. Consequently, the impact on the local economy is likely to be greater in these first months, certainly in terms of magnitude and probably in terms of velocity. An analysis such as this might prove useful when phasing relief in and out.

The examples above suggest that the overall impact of aid on conflicts may be overstated, but they do not imply that there is no impact at all. It is clear that, in countries like Afghanistan, relief aid to one degree or another is absorbed into the wider economy. Some proportion of relief resources is likely to reach warlords or rebel movements, whether through neglect or diversion. What these examples do demonstrate is the need for a clearer analysis of this process. By comparing the local value and velocity of aid to that of the overall resources a given group can access, it might be possible to approach an answer.

Roger Persichino has worked in Afghanistan and Sub-Saharan Africa as project manager and head of mission for several years. He recently carried out assessment missions in Zimbabwe, the northern Caucasus and Côte d’Ivoire. The Afghanistan and southern Sudan examples draw on personal experience and conversations with people familiar with both countries; for Somalia, material is drawn from meetings with NGOs and agency representatives in Nairobi, and from NGO situation reports.

References and further reading

S. Allix, La Petite Cuiller de Scheherazade (Paris: Hachette, 1999).

Paul Collier, Economic Causes of Civil Conflict and Their Implications for Policy (Washington DC: World Bank, 2000).

Paul Collier and Anke Hoeffler, Greed and Grievance in Civil War (Washington DC: World Bank, 2001).

Mark Duffield, Global Governance and the New Wars (London: Zed Books, 2001).

G. Garvy and M. R. Blyn, The Velocity of Money (New York: Federal Reserve Bank of New York, 1969).

François Jean and Jean-Christophe Rufin (eds), Economie des Guerres Civiles (Paris: Hachette, 1996).

Philippe Le Billon, The Political Economy of War: What Relief Agencies Need To Know, Network Paper 33 (London: HPN, 2000).

Joanna Macrae and Anthony Zwi (eds), War and Hunger: Rethinking International Responses to Complex Political Emergencies (London: Zed Books, 1994).

William Reno, Warlord Politics and African States (Boulder, CO: Lynne Rienner, 1998).