The perils of building democracy in Africa 

Can a get-out-the-vote effort decrease voter trust? A new study from Kenya says "yes."

January 21, 2016

Kenya elections poster

A poster in Kisumu, Kenya encourages citizens to “vote wisely” in the 2013 elections

New research on the impact of get-out-the-vote text messages sent to nearly two million potential Kenyan voters in the lead-up to the nation’s 2013 national elections found that texts have a measurable, if small, positive effect on voter turnout. But the messages—which were meant to instill confidence in Kenya’s fragile democracy—were also found to have had a negative effect on voter’s trust in the country’s electoral commission.

The study, “The perils of building democracy in Africa,” was conducted by MIT Sloan Associate Professor Tavneet Suri, MIT economist Benjamin Marx, and Harvard Business School Associate Professor Vincent Pons. It underscores the trade-offs between mobilizing voters and heightening their expectations about the integrity of elections in the developing world.

“On one hand, our results are encouraging: we show that sending text messages to potential voters is a quick, easy, and cost-effective way to reach people and increase turnout on Election Day,” says Suri. “But on the other hand, our results highlight the challenges of building trust in democracy—especially in developing countries that don’t have a lot of experience running democratic elections. When the institutions responsible for organizing and supervising elections raise voter’s expectations about the quality of the elections and then fail to fully deliver on their promises, voters feel distrustful and disillusioned.”

In 2007, after Kenya’s election commission overlooked irrefutable evidence of vote rigging to keep the ruling government in power, the country erupted into violence. The protests lasted for months and hundreds of people were killed.

The following year, Kenya adopted political reforms and a new constitution that, among other things, created the Independent Electoral and Boundaries Commission, which was tasked with creating a new register of voters across the country. Before the 2013 election, the commission purchased biometric voter registration kits, based on fingerprint technology, to mitigate identification issues at polling stations.

The commission also embarked on a voter education campaign. In the six days leading up to the election, the commission—working in collaboration with Suri’s team—sent eleven million texts to slightly less than two million prospective voters across Kenya. The messages were intended to rally voters and provided either basic encouragements to vote, background on the elective positions set up by the new constitution, or information on the electoral commission. 

“The IEBC was under intense public scrutiny during the electoral period,” says Suri. “After the deeply flawed and violent election of 2007, the new commission wanted to rebuild the trust of the people of Kenya. The text messages were its way of reaching out to voters in a gesture of honesty and openness.”

On Election Day, much of the expensive biometric equipment that the commission relied on to guarantee a transparent election failed, as did its system for electronically transmitting results. This shattered some of the high expectations the commission had set for a well-organized election.

Using official electoral results as well as independently collected survey data, the research team found that the text messages increased voter turnout by 1 to 2 percentage points. However, the intervention also decreased trust in the electoral commission by 5 percent.

“Getting voters to the polls and providing them education is essential to any democracy, but voters also need to have faith in the democratic process and the institutions at the helm of it,” says Suri. “Building that credibility and confidence is not easy, and promising and then failing to deliver can have detrimental effects on the very trust you are trying to build.”