Exit polls, surveys taken of voters as they leave polling centers, provide the most comprehensive information on the issues, perceptions and attitudes that matter to voters selecting candidates for office. They are used worldwide to help scholars, politicians, parties and citizens understand these important characteristics of the electorate.
First and foremost, exit polls seek to understand what determines why a person votes the way they do. What factors, including demographic characteristics and evaluations of government, contribute to the choice that citizens in a democracy make about their elected leaders? These matters are especially important in Kenya, where the transition to democracy has been long and hard-fought.
Many assumptions are frequently made about Kenyan voters; not just by the various political actors within the countries, but by scholars and policymakers from outside wanting to "promote democracy." They assert that Kenyan voters are intensely "tribal" - concerned only with expressing their ethnic identity, or driven by intense hatred of other groups - when voting for candidates from their communities. This suggests that Kenyan voters do not give due weight to considerations about government performance, and that elections are devoid of issues.
If this view is correct, we should remain skeptical about democracy's future.
These assertions, however, are long on rhetoric and argument, and short on actual evidence. Any observer of Kenyan elections knows that incumbent performance is always significant - which is why opposition candidates for president and parliament garner many votes. They also know that many issues defined this past campaign - corruption, constitutional revision, service provision, majimbo, employment and the economy were all debated.
Our exit poll is the set of data that can best help construct a profile of the Kenyan voter in this election. As scholars, we designed and implemented the most extensive exit poll in Kenya's history. On polling day, we scientifically sampled all types of Kenyans from every province, sixty-nine districts, and 179 constituencies. (Note that IRI is the sole funder, producer, and/or source of the exit poll.)
We asked voters a series of questions about their attitudes and evaluations of their local authorities, members of parliament, and the central government. We then asked which of the candidates they selected in the civic, parliamentary, and presidential races. By and large, voters select candidates by judging how well incumbents have performed in office, as well as where the candidates stand on the issues. Other important leadership traits matter as well. Ethnicity remains important for some voters, but it is not the only determinant of voting.
In our poll, President Kibaki won support based on positive evaluations of economic growth, his level of experience, and his performance in delivering free primary education and promise to do so with secondary schooling. The vote against Kibaki did not arise from overwhelming anti-Gikuyu sentiments in the electorate, but rather negative perceptions of his job in office.
His main challenger Raila Odinga won support from voters concerned with the high level of unemployment; his support of constitutional revision, majimbo, and fighting corruption; and the new ideas he would bring to government. The vote against Raila did not result from anti-Luo hatred against Raila, but rather positive perceptions of Kibaki's performance and the desire for him to continue in office.
Beyond the presidential race, many voters registered negative evaluations of parliamentary and local authority performance, and so voted out a significant number of former MPs and civic councilors.
Therefore, Kenyans are not simply tribal activists, but participate in elections with an eye to making important improvements to their own lives and that of their country. In this, Kenyans behave in much the same way as voters in older and more established democracies. From the point of view of citizens, therefore, democracy is alive and well in Kenya. But the post-election violence - even if perpetrated only by a few - demonstrates that ethnic animosity remains a problem. Politicians and other leaders need only rile a handful of angry supporters to cause mayhem and set back democracy.
Moreover, the strength of democracy is measured by the performance of institutions designed to guarantee free and fair democratic participation. The Electoral Commission of Kenya (ECK) failed the people when it proved so incompetent at properly counting ballots that its Chairman said he could not put faith in its results.
Because the validity of the ECK's results has been challenged from all sides, they are not the data that scholars, policymakers, commissions, or politicians should use when studying the election. The exit poll provides the only reliable check against the official results, and was itself immune to rigging or fraud. In the poll, Odinga won the nation-wide vote with 46.07% of the vote to Kibaki's 40.17%, a margin of nearly six points and outside the poll's margin of error. This is significantly at odds with what the ECK announced: a two point victory for Kibaki.
Because the ECK proved weak and ineffective and the parties placed no faith in Kenya's courts to fairly arbitrate electoral disputes, the international community played an over-sized and powerful role in negotiating a settlement. Power-sharing pleased powerful donors and foreign diplomats, but it was not what Kenyans voted for. Kenya's own institutions, most importantly the ECK, must be strengthened to improve the democracy that Kenyans demand. But commissions and agreements forced from the outside will not serve this purpose alone; the voters must continue to monitor and evaluate the performance of their leaders at the ballot box.
__________________________________________
|
TrackBack URI for this entry
Subscribe to this comment's feed
Show/Hide comments
EMOCRACY IN DIVIDED SOCIETIES’, Journal of Democracy, Vol 4(4), October 1993, pp. 18-38 
Show/Hide comment form
First, the good news. Your findings are very similar to Dercon's: in a pre-election poll conducted in early December 2007, ODM led 46.7 to 40.6 in the Presidential polling. In the Parliamentary polling, ODM led PNU 46 to 37.
Now, the bad news. Its relatively uncontroversial, I think, that PNU and PNU-friendly parties won the Parliamentary vote. It's also uncontroversial that the polls tightened considerably the closer the election came. If, as appears to have been the case, Kibaki-friendly parties closed the Parliamentary gap, why couldn't they have closed the Presidential?
Even worse news. You suggest - if I read you correctly - that opposition to Kibaki derived from negative perceptions of his performance. Unfortunately, Dercon finds that these were almost uniformly positive: 69% approved or strongly approved of Kibaki's performance (over the previous 12 months); 51% thought the country was doing well or very well under Kibaki compared to Moi; and even amongst likely ODM voters 47% approved of his performance. Indeed, even amongst ODM voters, fully 50% thought the country was doing better. If those numbers are accurate - and we have no reason to think otherwise - then we can eliminate dissapproval of Kibaki's performance as the driver of votes against him.
Further, we have the Bratton and Kimenyi's survey; their results seem to suggest that ethnicity remains perhaps the key driver of electoral behaviour in Kenya.