The independent resource on global security

Studying Somalia’s war economy from outer space

When we look at images of civil war such as the recent images from Syria, our gut response is that such destruction and suffering cannot and should not last. Surely war is a ‘means to an end’—and once that end is reached, civil society and the economy will rise from the ashes? Not necessarily so, says the literature on war economies: war can be 'economics by other means'. We test this proposition by using satellite images to reconstruct the unwritten economic history of the Somali civil war.

War can be seen as a rational economic activity. Warlords and insurgent movements use violence to extract rents: through looting and extortion or by charging taxes and ‘protection payments’ from traders and producers in their territories. International aid can be a significant revenue source for such ‘violence entrepreneurs’: it can be embezzled, diverted at roadblocks or directed to allied populations rather than the vulnerable and displaced. A ‘combat economy’ thus creates an elite of warlords, criminals and fighters with a direct interest in the continuation of war.

Do war economies really exist, or are these just isolated snapshots and impressions of jaded aid workers and frustrated peace campaigners? Who gains from civil wars? Do we prolong conflict by sending humanitarian aid into war zones? Can we even conduct an objective test of the war economy hypotheses? The problem is compounded by the fact that data collection usually ceases in the first weeks of war.

In a new multi-disciplinary paper published in the Journal of Peace Research Harris Makatsoris, Katerina Christopoulou and I overcome the problems of terrestrial data collection by using satellite images to reconstruct the economic history of the Somali civil war. US weather satellites take images of the earth every evening, and have collated the information as annual images of ‘stable nightlights’ since 1993. We can thus observe changes in electrical light output, which happens to be a very good proxy for disposable income in Somalia. Most electricity generation is private and small-scale, using diesel generators. Commercial generators offer pay-per-bulb electricity contracts to households in up-and-coming neighbourhoods. Indeed, we find that light output in Somalia is highly sensitive to imported food prices and local harvests, showing that at the margin people choose between lights and food.

The night-time images of Somalia tell interesting stories. First of all, Somalia is very, very dark at night. Only 15 urban centres were regularly visible in the period 1993–2009: coastal and pastoral populations simply did not develop well-lit public spaces and night-time economic activity. Second, major cities sometimes disappeared from the images altogether, while refugee camps outside them appeared, grew and shrank again. Third, light output collapsed in the anarchic south of Somalia but grew in the north. Finally, developments in the well-lit city centres do not correlate well with developments at the margins of the city.

Images of Somaliland—a self-declared autonomous Somali republic bordering Ethiopia—and especially its principal city, Hargeisa, as well as of Puntland—an autonomous region in the north of Somalia whose capital city, Garowe, was not visible in night-time satellite images until 1997—clearly document the economic recovery in the more stable regions of Somalia. In contrast, the depressing image of Mogadishu shows a profound economic collapse and the flight of the middle classes from the central areas to dimly-lit makeshift camps in the scrubland of the Afgooye Corridor.

So how is conflict intensity related to light output? We used a detailed data set of geo-coded conflict events in Somalia,  counted annual conflict events for each city and correlated this data with light output. We looked separately at total light output, light intensity in the brightest area of each city and the total lit area. The bright pixels in the centre of the images are our indicator for the fortunes of the elite who own businesses and residential properties in the city centres. The poor live on the margins of Somali cities—so if the lit area of the city expands it means that poorer areas gain access to electrical power.

We found that urban elites are resilient to both local conflict and to violence elsewhere in the country. If anything, elites in peaceful cities benefit from conflict in other areas which redirects diaspora investment to them. High intensity conflict in Mogadishu boosts total light output in cities hosting World Food Programme centres, as local economies respond positively to the influx of refugees, aid deliveries and the needs of aid providers. Only the poor unambiguously suffer during conflict: both local and remote conflicts significantly reduce light output from the cities’ margins.

So what did we learn about Somalia’s war economy? To begin with, Somalia’s urban poor would reap a substantial peace dividend from stable governance in Somalia. However, for the economic (and political) elite there would have been no economic incentive to negotiate an end to the conflict. Instead, elites in cities with a significant aid infrastructure would have gained from torpedoing peace settlements in southern Somalia.

The international community should consider these incentives in its state building efforts in Somalia and elsewhere. ‘Power-sharing’ agreements between elites need significant long-term fiscal transfers to those whose fortunes otherwise depend on the continuation of war. We need to ensure that humanitarian aid does not undermine incentives for peace-making. Democratization could be the foundation for long-term peace: it is the poor majority who reap the largest peace dividends and will demand peace at the ballot box.


This blog post is published as part of a collaborative partnership between SIPRI and Economists for Peace and Security (EPS).