Professor of Policy and Quantitative Methods
We expect voters to punish corrupt politicians at the ballot box, but a new study shows that Kenyans can overlook MP’s corruption and poor spending decisions if they share the same ethnic background. This suggests that elections alone are not sufficient for holding politicians to account.
“I am interested in understanding the main barriers for developing countries to become more advanced, focusing on governance, inequality, and conflict,” says Roxana Gutiérrez-Romero, Professor of Policy and Quantitative Methods at Queen Mary University of London.
Professor Gutiérrez-Romero has been studying Kenya’s main poverty reduction programme, the Constituency Development Fund.
“Earlier on in my career, I assessed the impact of the New Deal for Communities. This was a policy Tony Blair implemented in the UK giving two billion pounds to 39 deprived areas over ten years to spend however they wanted on the problem. I found the programme had very modest results.
“So, I was surprised the Kenyan government decided to run a similar programme giving $2 million dollars annually to MPs for them to decide in consultation with citizens how to address poverty in their constituencies,” Professor Gutiérrez-Romero explains.
When she first investigated this scheme in 2007, she discovered widespread corruption. For instance, according to media reports some money had been spent on projects that were clearly not aimed at reducing poverty, including building a golf course. So, in a follow-up study, Professor Gutiérrez-Romero decided to investigate how Kenyan voters react when they know how their MPs are misbehaving.
Voting intentions
Working with a team of 80 local researchers, she commissioned interviews with 1,200 people, representing all 290 Kenyan constituencies including areas so remote, they could only be reached by camel. Each interview was carried out in the interviewee’s mother tongue and used simple visual aids to communicate how their local MP spent the funds. The interviews were deliberately run after an election, so there was no risk of influencing voting behaviour.
Some of their findings have recently been published in the journal Political Behaviour.
Professor Gutiérrez-Romero continues: “We found that people do not know how their MPs spend this money. Once we informed them how it had been spent according to official audits, we asked them if they would vote for their MP again.
“In some cases, we would remind them which party their MP belonged to and in these cases, people changed their mind and said they would not vote for the same MP again.
“In other cases, we deliberately mentioned the MPs surname, because in Kenya, this reveals their ethnicity, for example whether they are Luo or Kikuyu. When we did this and voters realised that their MP was from the same ethnic group as themselves, they were very forgiving and said they would vote for the same MP again.
Defying expectations
“This goes against everything we think we know about voter behaviour. According to the literature in political science, it’s assumed people vote for their co-ethnic because their co-ethnic will know how they want services to be delivered.
“I was able to use real data to show voters how their co-ethnic MP misspent funds, yet they said they would still vote for that MP. The findings suggest that some voters think, ‘I’d rather have a corrupt guy who will at some point protect me rather than a corrupt guy who will protect someone else; when they’re one of my own, I know what to expect’.
“Several theoretical studies predict that once you update people on how their politicians behave, they will react rationally and that’s not the case. These imaginary elections used to study political behaviour can be useful, but they have limitations.
“What matters is how voters assess candidates in real elections and how they update their beliefs once they’ve been shown evidence on how their MPs used poverty funds. People have attachments and that makes them deviate from our expectations.”
A new approach
Professor Gutiérrez-Romero adds: “For me, a key message from this research is don’t give $2 million to MPs and ask them how to spend it.
“There’s an incentive to spend money on more visible projects, like repainting a town square, but that’s a waste of money if it doesn’t lift people out of poverty. At the moment, there is not much cooperation between MPs, and spending money locally can be very divisive. Perhaps areas could coordinate to jointly build projects for the benefit of the wider region.
“This is Kenya’s major anti-poverty programme. Asking people what they want is important, but to lift an entire community out of poverty needs additional analysis, and that has been missing. I think a more deliberative approach is needed.
“Since we did this research, I’m pleased to say that Kenya has reformed the Constituency Development Fund. MPs simply monitor the programme; they do not decide how the funds are used. This has removed the potential for misuse.”