Skip to main content Skip to navigation

Guns 'n' roses


Research from Kenya’s flower fields reveals the economic toll that ethnic violence takes on businesses.

Christopher Ksoll,Rocco Macchiavello and Ameet Marjariaconsider the policy implications for violence-plagued regions.


Elections in Africa are often dangerous times. Two out of 11 African presidential elections in 2007 and 2008 generated large-scale violence and in the others, the risk of violence was a focus of media coverage. Adding to the human tragedies unfolding in these countries are the economic consequences on businesses, a little-studied subject.

Our research examines how a short, intense period of violence affected firms in Kenya’s flower industry. Our results have broad implications, offering new insights into connections among firms, industry associations, and international trade flows in the face of sporadic ethnic violence. The research also sheds new light on policies and practices that may reduce the violence or mitigate its negative economic effects.

Relatively few studies have focused on how ethnic violence affects firms, largely because of the enormous research challenges inherent in such situations. Certain characteristics of the situation that unfolded in Kenya allowed us to overcome some of the most daunting impediments. The sporadic locations of the violence and the detailed record-keeping that is a hallmark of its export-oriented flower industry allowed us unusual opportunities to gain insight about the economic dynamics at work.

With its location on the equator, perpetual sunshine and variations of altitude, Kenya is endowed with many of the natural conditions necessary for growing flowers. In slightly more than a decade, the country has become one of the world’s largest flower exporters, overtaking such traditional leaders as Israel, Colombia and Ecuador. The flower industry is now one of Kenya’s largest foreign currency earners, alongside tourism and tea.

Jobs created by flower firms are valuable and seem to have discouraged participation in the ethnic violence that ravaged Kenya.

The industry is labour-intensive and employs mostly low-educated women in rural areas. Flowers are fragile and highly perishable, and, as a result, post-harvest care is a key determinant of quality. Workers, therefore, receive significant training in harvesting, handling, grading and packing, and they acquire skills that are difficult to replace in the short run. Flowers are exported from Kenya either through auctions in the Netherlands or through direct sales to wholesalers and specialist importers.

Kenya’s fourth multi-party general elections were held in late December 2007, and three days later, the incumbent president Mwai Kibaki was declared the winner over Raila Odinga, the opposition candidate who had been leading in published polls. Within minutes of Kibaki’s swearing in that evening, a political and humanitarian crisis erupted. Targeted ethnic violence began then and ended in late February 2008. In its wake, an estimated 1,200 people were killed and more than 300,000 others were displaced. Financial losses to the economy reached approximately £145 million, around 1 percent of gross domestic product.

The violence took place in certain regions where flower businesses are located and spared others – thus, opening a window of opportunity for our empirical research. One of the most difficult aspects confronting empirical studies in this situation is to provide what is called the counterfactual, in our case, a valid assessment of what would have happened to the firms affected by violence if the violence had not taken place. The geography of the Kenyan violence unwittingly offered the perfect counterfactual.

Another empirical challenge is gathering detailed information on the operations of firms exposed to violent conflict. Again, the features of the Kenyan flower industry provided a way to overcome this hurdle. Flowers in Kenya are produced by about 120 established exporters exclusively for foreign markets. As a result, administrative records for all firms in the industry include extensive daily data on production and sales. In addition, shortly after the violence, we were able to conduct surveys in the field with firms about the channels through which the fighting affected the firms’ operations. As a result, our case study allows us to examine the microeconomic effects of the outbreaks of violence with a remarkable degree of detail.

During a two-week episode of intense violence, export volumes and revenues of firms located in affected regions dropped by 38 percent relative to comparable firms in regions not affected. This effect can be separated into two distinct components, a 31 percent drop in export volumes, mostly associated with missing workers, and a 9 percent drop in the likelihood of exporting in a given day, mostly associated with transport problems.

These average figures, however, conceal substantial differences in both firms’ exposure and response to the violence. In particular, large firms, firms with stable contractual relationships in export markets, and firms affiliated with the industry association were better able to handle the difficult situation. They were able to take a wide variety of measures – some cheap but others costly – such as hiring extra security for transportation and setting up camps and temporary accommodation for workers, to encourage them to come to work rather than stay on guard at home, for instance. On average, operating costs for these businesses grew by 16 percent.

Firms affiliated with the industry association suffered lower reductions in export volumes, presumably because the association provided the necessary help in coordinating transportation issues.

There is no evidence that foreign-owned firms or firms more closely connected to local politicians suffered differentially because of the violence.

The firms best able to handle the crisis were large, export-oriented and affiliated with an industry association that could provide help.

The jobs created by flower firms are valuable and seemingly discouraged participation in the violence. Kenya is an ethnically fragmented country, and the violence in the wake of the elections pitted members of certain tribes against others. Most flower firms have ethnically diverse workforces, composed of labourers from tribes that were in conflict with one another. Yet no violence was directed towards flower firms, and no episode of violence was recorded to have happened on firms’ premises.

The export-oriented nature of the industry further contributed to stabilising the situation, because firms sought to find ways to fulfil contractual obligations with foreign buyers. As a result, policies directed at upgrading agricultural products towards commercial exports might have beneficial side effects in mitigating the risk and consequences of violence.

Publication details

This article summarizes “The Effect of Ethnic Violence on an Export-Oriented Industry” a working paper by Christopher Ksoll, Rocco Macchiavello and Ameet Morjaria. The paper can be accessed at:

The authors

Christopher Ksoll is a research officer at the Centre for the Study of African Economies at Oxford University. Rocco Macchiavello is assistant professor of economics at Warwick University, a fellow of the Centre for Economic Policy Research, and an affiliate of the Bureau for Research and Analysis of Economic Development. Ameet Marjaria is completing a PhD at the London School of Economics.