As the Kenyan election campaign hits top gear, the pros and cons are becoming clear for all to see.
On the positive side, the presence of a strong opposition has forced the Jubilee government to defend its record, and means that Kenyans have a real choice to make when they head to the ballot box on August 8.
But there is also a negative aspect to the campaign, which has once again seen rival leaders trading accusations and insults, alongside a rise in hate speech and the use of ethnic stereotypes.
I have spent much of the last 10 years defending the potential and performance of democracy in Africa against pessimists who argue that it has not, and cannot, work.
Against this, I point out the reintroduction of multiparty politics has enabled millions of Africans to remove corrupt and dysfunctional governments, from the early defeat of authoritarian regimes in Benin and Zambia to the unexpected opposition victories in Gambia and Nigeria in the last couple of years.
However, the transformative potential of electoral politics should not blind us to the challenges elections present to political order and national stability.
Even committed democrats must recognise the “dark side” of elections — if only to work out how it can be better managed.
Even in established democracies, elections are moments of great national stress.
They dominate media coverage and public conversation, lead to more polemical columns in newspapers, and cause Facebook friends to fall out.
The stress elections generate is rooted in uncertainty.
In established democracies, voters fret about which candidate will win and what that will mean for their lives and values.
Just think back to the tears shed by Democrats after Republican Donald Trump’s election victory in the US last November to see how keenly citizens sometimes feel the defeat of their parties.
In new democracies, the stakes are even higher. Winner-takes-all politics means the cost of losing is greater, and the fragility of political institutions means the result of the election is not the only thing to worry about.
Right now, for example, many Kenyans are thinking about whether President Uhuru Kenyatta of Jubilee will hold on to defeat Mr Raila Odinga of the National Super Alliance; whether Mr Odinga will accept defeat if he loses; whether the opposition leader will call his supporters to the streets if he rejects the result; and, whether that will lead to widespread political violence.
Most of the Kenyans I know believe passionately in their country, and in the peaceful inclinations of their fellow citizens, but this has not stopped them from fretting.
As a Nairobi-based friend recently put it, “if you have high blood pressure, elections can be bad for you”.
The physical evidence of these stresses can be seen in the thousands of people who travel “home” for elections.
Home, in this context, does not necessarily mean where you live or where you work but where you feel safe — that is, where your own community is the overwhelming majority.
To better understand why this happens, and what it tells us about the broader impact of elections on national identity, we need to look at the impact of elections on society writ large.
One of the most obvious problems associated with elections is they often harm the relationship between different ethnic groups or classes.
In-between elections, citizens are less frequently reminded of the issues that divide them, which allow a stronger sense of national identity to come to the fore.
By contrast, in the heat of the battle for hearts and minds, leaders and activists say and do things that exacerbate inter-communal tensions.
Age-old stereotypes are dragged out of the cupboard for another airing.
Familiar accusations are resuscitated, and may even be exaggerated in a bid for shock value.
In the Kenyan context, these tropes are well known. While many candidates abstain from this kind of mobilisation, some have turned it into an art form.
For example, unscrupulous opposition activists try and whip up anti-Kikuyu feeling, accusing the community of monopolising economic and political opportunities.
For their part, irresponsible government figures will attack Mr Odinga on the basis of his identity and culture, depicting him as an “uncircumcised boy” unfit to lead the country.
At the same time, some of the groups that make up Nasa have been described as backwards and blamed for their own economic difficulties.
Although these stereotypes are so familiar as to have reached the status of clichés, the way they are repeated during campaigns is damaging.
For example, research by Benn Eifert and his colleagues at the University of California, Berkeley, has found that elections in Africa make people feel more “ethnic”.
By comparing how people choose to identify themselves in surveys conducted close to elections as opposed to those carried out in the middle of the president’s term in office, they are able to tell how people’s feelings about their own identities vary over time.
Their findings are compelling: The closer to an election a survey is held, the more likely it is those interviewed will emphasise the importance of their ethnic identity as opposed to, say, their national or their professional identity. This is important, because it means elections serve to reinforce the kind of divisions that can — if they are mishandled — lead to ethnic clashes.
The other big loser during election campaigns is public trust in political institutions.
As opposition leaders criticise the performance of the Independent Boundaries and Electoral Commission (IEBC) and call into question the capacity of the courts to deliver justice — in some cases with good reason — public confidence in the state takes a knock.
Of course, this is not the case everywhere. In the continent’s high quality democracies, in which elections are well managed and lead to consensual outcomes, elections need not have this impact.
But in countries in which rival parties have yet to fully agree on how elections should be run, and in which the outcome of previous elections remains controversial, elections undermine the notion — essential in a successful democracy — that the state is neutral and acts in the interests of all citizens.
Public attitudes to the IEBC demonstrate this point well. In a recent nationally representative opinion poll of 1,100 Kenyans conducted in June, 37 per cent of Kenyans reported that they trust the commission “a lot”.
A further 27 per cent of people said they “somewhat trust” the IEBC, while 33 per cent professed “little” or “no” trust at all.
These headline figures do not look too bad at first glance, but these averages mask considerable — and worrying — variation between opposition and government supporters.
While a strong majority of Jubilee supporters (64 per cent) had the highest level of confidence in the IEBC, only a small minority of Nasa supporters (18 per cent) felt the same.
Similar patterns can be found on other election related issues.
In the same poll, 91 per cent of Jubilee supporters said that they were “completely sure” that their vote was secret, but this was true of just 79 per cent of opposition supporters.
These patterns suggest the efforts of opposition leaders to call into question the credentials of key political institutions, along with the failure of these institutions to effectively respond, has undermined the confidence of their supporters in the electoral process.
In turn, this makes it more likely that Nasa voters will disbelieve the result of the election if their candidates are unsuccessful, increasing the risk of post-election unrest.
Of course, the broader impact of elections cannot be separated from the result and what comes next.
On the one hand, a disputed election can lead to civil unrest, further undermining inter-communal and institutional trust.
When this happens, as it did in Kenya in 2007/2008, the stakes are raised even higher for future elections.
On the other hand, a good quality election can help to overcome past controversies.
The New Progressive Party in Ghana stopped complaining about the 2012 election after it won power in 2016.
Similarly, the All Progressives Congress in Nigeria changed its tune on the quality of the electoral system after it defeated the once dominant People’s Democratic Party in 2015.
The positive effects of a good quality election is not limited to the political elite.
In 2004, the American political scientist Michael Bratton demonstrated a transfer of power has a significant impact on public support for democracy.
Using the same Afrobarometer data employed by Eifert et al, Bratton shows in many countries the defeat of the ruling party “refreshes” citizens’ “commitment to democracy”.
In other words, one of the best antidotes for a series of bad elections is a good one.
This is why it is worth persevering with elections, despite the multiple difficulties they cause, and why a good quality poll is so important on August 8.
Nic Cheeseman (Twitter: @fromagehomme) is Professor of Democracy at the University of Birmingham, UK