For decades urban residents in many Africa countries have shown their dissatisfaction with their governments. Food price riots that rocked Guinea, Mauritania, and Senegal in 2007-8, and Uganda's "Walk to Work" protests in 2011 were testament to this. As were the protests that rippled through Sudanese cities in 2018.
But, hostility towards incumbent governments is usually subtler. Riots and protests are the tip of the iceberg. New evidence from public opinion data clarifies the extent of this urban dissatisfaction.
In my new book Rural Democracy: Elections and Development in Africa, I use survey data from Afrobarometer, the independent African research network, from 28 countries over ten years (2005-2015) to document the extent of urban hostility towards incumbent governments across Africa. The picture is not rosy.
Throughout the continent, urban residents are significantly less supportive of incumbent governments than rural residents. This varies quite dramatically with considerable variation between countries. But, on average, urbanites are between 5 and 7 percentage points less likely to say that they would vote for the incumbent.
Differences in satisfaction with democracy are similar. Urbanites are at least 5 percentage points less likely to report being satisfied with democracy in their country, on average.
These findings give a strong sense of the average level of urban hostility towards incumbent governments. While the negative sentiment is widespread, there are sizeable differences in its extent. For example, urbanites in Burundi are 18 percentage points less likely to support incumbents. And they are 22 percentage points less likely to be satisfied with democracy. In Malawi and Tanzania, urbanites are around 10 percentage points less likely to support incumbents, and 5 percentage points less likely to report satisfaction with democracy.
At the other end of the spectrum lie Botswana and Zambia. These cases see no significant differences in support for incumbent leaders or satisfaction with democracy across urban and rural areas.
A major benefit of comparative, cross-national research is the ability to look across cases for systematic patterns, as I have done in the book. This enables the development of generalised explanations, to make sense of the observed variation.
Why is there trouble in the cities?
Trouble in the cities
Part of the greater rural support for incumbents is explained by structural differences across urban and rural contexts. For example, lower political opposition in the countryside means incumbents face less competition.
In rural areas there is also more clientelism - the exchange of private goods and services for political support. This matters because access to state resources gives incumbents an advantage in dispensing patronage. And traditional authority often remains stronger in the countryside. This means incumbents can also engage local authority figures to deliver votes.
These factors help make some sense of the difference between urban and rural areas. But they leave much unaccounted for. They cannot explain, for example, why incumbents risk hostility in urban areas. Nor do they account for the wide variation in urban dissatisfaction across countries.
One reason, is that most African countries still have majority rural populations. Urban discontent stems from incumbents implementing policies that favour rural areas to win elections.
Looking after rural interests and paying less attention to the demands of urban voters is, therefore, a feature of democracy in predominantly rural countries.
It wasn't always like this. Under authoritarian rule and in the absence of meaningful electoral competition, the main threat to rulers came from coups. These were easier to coordinate in urban areas. So, to prevent urban unrest, authoritarian leaders tended to be biased towards urban areas.
But, with the reintroduction of multiparty elections the discrepancies in political power between urban and rural areas have been largely equalised by the ballot box, leading to rural bias in policy-making.
A change in weighting
There are signs that this is beginning to change as countries urbanise. This decreases the incentive to prioritise rural voters. That means incumbents are likely to balance pro-rural and pro-urban policies more evenly.
This reduction in rural bias in line with urbanisation should diminish urban dissatisfaction, which is exactly what I find.
Notwithstanding some outliers, differences in support for incumbents and satisfaction with democracy across urban and rural areas is greatest in the least urbanised countries. The size of country-specific coefficients estimating the impact of urban residence on incumbent support and democratic satisfaction are negatively and significantly related to urbanisation.
The relationship becomes clearer still when urbanisation is interacted with the urban/rural indicator in estimates pooling data across all countries. These estimates provide robust and compelling evidence that urban dissatisfaction across Africa is mitigated by urbanisation. The more urban a society, the less unhappy urban citizens are. The more rural it is, the higher the levels of urban dissatisfaction.
These findings offer strong reasons to believe that trouble in Africa's cities stems from the fact that electoral competition drives leaders to be biased towards rural areas.
This provides important new insights into the dynamics of electoral competition in Africa. Not only does it explain widespread urban dissatisfaction, but it also offers an alternative to the received wisdom that elections are predominantly contests in corruption, clientelism and ethnic mobilisation.
While these are prevalent, elections across Africa are also being fought over policies and which types of voters they favour. To the extent that democratic accountability structures operate this way, democracy in Africa works.