Cleaning up Dirty Elections – what works?

October 15, 2009

     By Duncan Green     

The Centre for the Study of African Economies in Oxford (home to Paul Collier, among others) is putting out some fascinating two pagers on its work, including two recent papers on ‘dirty elections’.

election violence in KenyaIn ‘Cleaning up Dirty Elections’ Paul Collier and Anke Hoeffler go to work  on a new data set spanning nearly 30 years and 155 countries (suggesting that the CSAE is expanding its empire beyond Africa) and find a mixture of ‘well duh’ and more interesting connections.

In the ‘well, duh’ category are ‘Using dirty tactics during elections helps politicians that are already in office’ and ‘Small, poor but resource-rich countries are more prone to dirty elections.’

But more interesting are:

‘Dirty elections are bad for economic growth by skewing politicians’ incentives towards pursuing bad policies rather than good ones;

Checks and balances (Term limits, a free press and constitutional checks and balances are the most effective) reduce the incentives to cheat and implement bad policies.

[Whereas] international aid has no clear effect on the quality of elections, unless there are effective checks and balances.’

Another CSAE paper analyses the impact of an ActionAid International campaign against voter intimidation in Nigeria in the 2007 elections, in which election violence in Nigeriaover 300 people were killed. Using violence to intimidate voters was the strategy used by the opposition politicians (incumbent politicians tended to use vote buying and fraud).

ActionAid’s campaign consisted of holding town meetings, street theatre productions and the distribution of leaflets in six states. The CSAE ran household surveys and contracted local journalists in each observed location to keep diaries of local violent events. To allow a clear attribution of the results to the AAIN campaign, comparisons were made with similar locations that were not part of the anti-violence campaign.

Main findings were that in areas targeted by the campaign:

Less violence occurred;

Violent politicians got fewer votes, because more of their former supporters abstained;

Voter turnout increased by 10%.

The campaign was especially effective with those people who were less locally integrated because they were poor or working outside the district. CSAE concludes that this group was less likely to benefit from local political deals and were therefore more receptive to the campaign messages (although I would guess that there are other possible explanations, like this group receiving less information from media and other sources, and so more influenced by a campaign).

The CSAE has not yet set up an email subscription system, but until they do, if you want to receive updates on their publications, contact Karin Loudon at

October 15, 2009
Duncan Green