The recent wave of South African university student protests (over university fees and other grievances) comes on the heels of more than a decade of widespread protests across South Africa over basic service delivery. Organizing on the streets, carrying placards and sometimes engaging in violence, South African citizens have stood up in increasing numbers to challenge the ruling African National Congress (ANC). They say that they are disappointed with the government’s failure to provide houses, water, toilets and other services, and have often issued threats to withhold their votes in upcoming elections.

Such overt challenges signify the honeymoon is over for Mandela’s liberation party. When the apartheid regime, one of the most infamous governments of the 20th century, fell in April 1994, the relatively peaceful transition ushered in a period of optimism. On the back of these hopes, the ANC ruled with an emphatic political mandate for more than 20 years. Mandela and his successors promised “a better life for all.” But in recent years, protesters, ordinary citizens and the media have argued with increasingly louder cries that this promise has not been realized.

South Africa remains one of the most unequal countries on the planet, and millions of citizens remain in dire poverty, without access to the most basic services. Yet it would be remiss to ignore some of the government’s exceptional achievements in service provision. Our calculations using census data from StatsSA – a body comparable to the U.S. Census Bureau – show that between 2001 and 2011, around 1.9 million new households received piped water, 1 million received flushing toilets, and 660,000 received regular refuse collection. Access to electricity services – albeit with a national grid that is regularly incapable of providing reliable electricity – has become nearly universal.

As political scientists and close observers of South African politics and development, we saw these uneven accomplishments (and the ensuing public challenges) as an opportunity to reflect on one of the central tenets of democratic governance: that voters reward the incumbent for good service, and punish it for bad.

On the face of it, the South African case raised the possibility that voters could be collectively holding the government to account. We wondered if the variation in service provision along with variation in demonstrable frustration with the ruling party would provide some clear evidence that democracy could work as democratic governance analysts expected.

In other words, were the people taking to the streets the citizens who had been left behind in the ramp-up of public service provision?

Armed with access to high-quality data, we examined a related question: Do citizens with improved service provision reward the ruling party with their votes?

We carried out a comprehensive study of how changes in service provision (water, toilets, and refuse collection) affected changes in ANC vote share in both local and national elections between 2001 and 2011. The results are striking. We find exactly the opposite of what we – or more precisely what fundamental theories of electoral democracy – had predicted.

We find that increases in service access do not induce increases in the ANC’s vote share. In fact, quite the opposite occurs. In communities which show large increases in service provision (piped water, flushing toilets, refuse collection), changes in ANC vote share are actually lower, often even negative, relative to those areas which saw smaller or no gain in services.

These findings are also mirrored in nationally representative individual-level survey data, collected by Afrobarometer every few years. We focused our analysis on black South Africans, because this group was systematically denied services under the old regime, and had the greatest hope for improvements under the ANC.

We find that those who had access to various basic services reported being less likely to vote for the ANC in future elections–despite the fact that individuals with such services did report enjoying a better quality of life.

In other words, individuals with access to basic services appear to be happier with their living conditions, but they are less likely to want to keep the ANC in power, at least compared with those still living without services. Further, our findings are not driven by things like education, income, wealth, or geographic location.

Why do voters in South Africa not reward the ANC? After we ruled out the results being driven by the ANC strategically providing services to areas that support them, we considered other potential explanations for our surprising finding.

Two different pathways through which increased service provision fails to induce support find support in our data, but we hasten to point out that these particular findings are preliminary, and demand further investigation.

First, increases in service delivery appear to heighten citizens’ awareness of, and exposure to, corruption. Corruption is a salient political issue in most developing democracies, and so increased exposure to corruption may shift voting behavior. One could imagine that those places in which there is the greatest improvement in services are also those places in which local politicians self-enrich the most.

Second, increases in service delivery appear to change voter expectations, ratcheting them upward. Once voters are provided with basic services, they may alter their expectations and demands of government, seeking out alternative parties. This could occur either through changing needs and desires, or through revised understandings of government capacity.

Our striking results find interesting parallel in the student protests that disrupted South African politics in the last weeks of October. The students, along with the others protesting the government, represent a very particular demographic: young, educated people.

This suggests that political change in South Africa may come, but not as sometimes portrayed or assumed. Those who have benefited least from the apartheid transition may continue to “double down” on the hopes that the ANC will finally deliver to them. Those who have benefited most seem most likely to drive political change.

While these findings are surprising, we don’t interpret them as bad news for democratic practice. Citizens are making choices and taking actions in the electoral arena based on their experiences and their hopes for the future, just not in the way we had always assumed. Indeed, our results suggest that political scientists and commentators should continue to tread carefully when assuming what voters really want, and why.

Daniel de Kadt is a PhD candidate in political science at the Massachusetts Institute of Technology (MIT). You can follow him on Twitter at @dandekadt. Evan Lieberman is the Total Chair in Contemporary African Politics and professor of political science, also at MIT. You can follow him on Twitter at @evlieb.

Related posts: