Do South Africans value democracy?


back to issue

FOR democracy, these are perhaps the best and worst of times. The best, because formal political freedom is more pervasive than ever. The worst, because democracy’s relative triumph has been won at the cost of its ‘hollowing out’, both practically and in the value which is placed upon it.

New democracies seem largely unable to translate voter preferences into policy.1 This article will, however, focus not on this but on the degree to which democracy has, among elites rather than citizens, come to be seen less as a value in itself, a source of participation and liberty, and more as an instrument of economic growth and administrative effectiveness.

It will take as a case study a country whose elite would, given its history, be expected to take democracy as a value more seriously than most: South Africa. It will, however, argue that, in contrast to citizens, its elite is adopting a particularly instrumental view of the society’s new-found right to govern itself. And it will illustrate this through an analysis of the country’s second universal franchise election in June 1999.

Elections are not solely a means of counting public preferences. Many citizens may see an election as the only occasion on which their choices are stamped on the political process. Voting is an act of giving voice and thus of asserting the political self, of expressing identity and autonomy; one analysis suggests it is also a crucial act of commitment to democracy and its values, capable of developing ‘positive, democratic character traits such as community mindedness, political self-competence, and satisfaction with decision-making structures, institutions and outputs.’2 That citizens should be able to exercise their vote, and derive satisfaction from doing so, is more important to democracy’s health than for whom they elect.

This point is particularly important to South Africa’s new democracy for, as citizen behaviour in 1994 showed,3 the vote takes on added meaning in a society in which a racial minority monopolised it. Just as the denial of the vote was a symbol of exclusion, its achievement became a sign of restored dignity.

This may seem obvious. But, during the negotiations which produced the 1993 settlement ending apartheid, negotiators representing the African National Congress, the present government which led the fight against racial domination, insisted that only special measures would induce many black South Africans to vote because the violence of the past few years had made them fearful of the risk.4



Understanding voting in this way – as an instrumental exercise in which citizens weigh the uncertain benefits of casting a ballot against the sure cost of doing so – is a hallowed tenet of rational choice theory5 which has failed repeatedly to explain actual behaviour.6 As later events showed, its application in South Africa borders on the absurd.

This is so not only because the memory of a racial franchise is a spur to voter enthusiasm. The ANC’s dominance at the polls, and the consequent perception that the outcome of elections is certain, would make voting irrational were it not spurred by factors other than cost-benefit calculations. Prime among these is identity.7 South African parties are defined by, and draw their support from, identities – race, language and religion primary among them. Casting a ballot is primarily not an instrumental calculation but an expression of who a citizen is. And people will go to great lengths to express who they are.

In principle, South Africans’ enthusiasm for voting gives it an important resource not always available to embryonic democratic systems – a degree of citizen commitment to democracy.

And, while some of the points made here about the first election may seem obvious, they apply equally as much to the second. If, for some commentators, the second elections are less important than founding ballots, this is not necessarily the perception of African citizens: a survey of 15 African second elections reveals that almost half recorded higher voter turnout than the founding election,8 confirming that citizens’ participatory impulses do not disappear when the first election ends.

This background provides us a prism through which to view the 1999 election and the degree to which it provided a vehicle for, or obstacle to, citizen support for democracy as a value.



The framing of the electoral rules and conduct of the campaign reveal that the elite’s instrumental perception of the purpose of elections, as well as of citizens’ propensity to vote – have not changed fundamentally since 1993.

On the first score, the ANC concluded that preventing electoral fraud by insisting that voters could only cast a ballot if they acquired an identity document which contained an electronic bar-code outweighed extending the franchise to as many eligible adults as possible, since a significant section of the population lacked the documents and were unwilling or unable to jump through the bureaucratic hoops required to obtain them.

There is no firm evidence for the claim of white-led opposition parties that the identity document stipulation disadvantaged their voters disproportionately; a study found that several hundred thousand voters in ANC strongholds also lacked the required ID. Rather, the choice appears to have been underpinned by the assumption that the chief purpose of the election was to produce a technically unassailable result and that this merited curtailing the number of adults who would enjoy the democratic opportunity which voting provides.



A similar lack of enthusiasm for extending participation as widely as possible is suggested by the attitude to eligibility of the Independent Electoral Commission which was responsible for administering the poll. It adopted a US-style approach which placed the burden of enrolling as a voter on the citizen. In the US, registration criteria are arguably as responsible for low voter turnout as indifference or the instrumental calculations of rational choice theory.9 It might be argued that, because voting is an act of democratic identification, the democratic state, if it wishes to ensure its health, has a core responsibility to ensure that this opportunity is available to all who wish to take advantage of it.10

But if citizens are enthusiastic about voting, will they not be willing to register? If queues at official offices are a guide, many were prepared to do just that. But this implies that all citizens possess the capacity to take advantage of these opportunities. While officialdom did make some effort to lengthen office hours to allow citizens to obtain the required ID, it is unlikely that everyone who needed these documents was able to access them. And officials do not seem to have found a way of conveying information on registration to many citizens – a survey found that upto 45% of eligible adults who did not register believed they were still entitled to vote.



Nor did the IEC do all it could to expand the corps of registered voters, even within its own rules. Much of its activities were based on the assumption that it was voters’ responsibility to approach it for the vote, not its duty to seek voters out. The result was registration of some 75%, according to the IEC, but which may have been as low as 70%.11 So almost a-third of the adult population may have been deprived of their right to vote. Since even the most pessimistic survey on voting intentions found that some 75% planned to vote,12 at least 5% of citizens who wished to exercise the franchise may have been unable to do so.

This level of exclusion is not nearly high enough to question the legitimacy of the election. Contrary to the apparent calculations of the politicians and the IEC, the real loss lay in the denial of democratic participation and the implied assumption that participation is less of a value than factors such as cost or administrative neatness.



The elite’s understanding of motives for participation was suggested by the election’s central issue: whether the ANC would win two-thirds of the vote, enabling it to change the constitution. While white-led opposition parties used this prospect to mobilise their voters, the ANC started the ball rolling by conjuring up images of the liberatory effects of the magic number.13

But why did it and the opposition use the size of the majority as a theme since, during the campaign, the ANC showed little interest in constitutional change? Both believed that since the election winner was not in doubt, only focusing on the size of the majority would get out its vote. As part confirmation, a senior ANC official exhorted researchers not to publish an analysis which indicated that it would fare better than was then predicted because ‘we will then not be able to motivate our people.’14 So the two-thirds debate was the product of a community of interest between governing party and opposition in manufacturing an issue which would encourage voters who firmly expected their party to lose or win to turn up at the polls.

This suggests a common perception that, at least to a degree, the rational choice theory of voter behaviour was accurate. By setting a much higher threshold for ‘deciding’ the election – two-thirds rather than 50% – the parties sought to increase the degree to which voters perceived a concrete benefit from voting so that they would be more willing to incur its costs.

But is this understanding of South African voters as people who weigh their decision on whether or not to vote on utility-maximising criteria valid? There is anecdotal evidence of voters who did register and vote because they wished to affect the ‘two-thirds’ result but it is hardly enough to support or refute the argument. We do, however, have a test of the degree to which the perceived ability to influence the result determined propensity to vote – the provincial percentage polls. If voters are more likely to vote if they believe their choice will determine the result, we would expect higher polls in provinces where the outcome of the contest was in doubt. But there was little difference between provincial percentage polls; the highest were recorded in provinces where the victory margin was most lopsided, the lowest in a closely fought province.

The test is hardly definitive. Voters may not care whether their ballot influences provincial outcomes if they worry only about the national result, and rational choice theorists could protest, with justification, that it is irrelevant to the theory.15 But our purpose is to test the claim that South Africans vote on instrumental criteria: the only data available from the 1999 poll suggests that they do not, and that citizens’ voting behaviour is influenced far more by democratic commitment and enthusiasm than the elites allow.



Given the analysis sketched thus far, it is no surprise that elite assumptions prior to both elections assumed a citizenry far less attuned to democracy’s benefits than their betters in the media, politics, academe and the NGO movement.

This has many symptoms, of which enthusiasm for ‘voter education’ is an example. The assumption that voters need to be ‘educated’ implies that they are unable to cast a ballot without aid, which was refuted by behaviour in both elections, or are unable to discern their interests without help, which has also been empirically refuted.16



Another is the claim that voters are inherently apathetic. This takes different forms, depending on whether voters are in the racial minority or majority. For the former, ‘apathy’ is said to stem from disenchantment with majority rule,17 for the latter from reaction to inadequate government delivery of public goods, a claim which appears to lack any evidence at all. The turnout on election day, 1999, would appear to contradict both claims. The official percentage poll was 89% and there is no evidence that participation among racial minorities was lower: on the contrary, Indian voters who in survey evidence are said to feel most alienated by the post-apartheid polity, are said to have turned out in particularly large numbers.18 This does not mean that these citizens are not disenchanted; merely that, if they are, they choose to express their grievance at the polls.

It is also worth mentioning that, while the inconvenience to which voters were subjected in 1999 rarely paralleled the 1994 experience, there were polling stations in which voters were subject to substantial discomfort. And, even at stations which experienced less obvious logistical problems, the queues were often long enough to persuade anyone in them whose democratic commitment was tenuous, to abandon their quest.

Something is also worth saying about the ethos at polling stations. As in 1994, the experience of voting appeared to instil a sense of camaraderie and mutual regard, expressed in some cases in acts of generosity not always evident between elections. Despite a bad-tempered and sometimes violent campaign, polling day violence was largely absent. Many citizens may derive sufficient satisfaction from voting and enough of a sense of identification with the society to induce an unusually high level of ‘civic’ behaviour.19



The levels of participation may, therefore, confirm that the degree of citizen commitment to democracy may be much higher than presumed by elites. But it may also say something important about the preconditions for effective governance.

South Africa is often portrayed as an unruly society in which citizens resist meeting their side of the ‘social contract’ with the democratic state. There is much evidence to support this, such as high crime and poor payment levels for public services. But the election invites reassessment of the assumption that South Africans are beyond the reach of the state unless it uses force or better administration to ‘cage’ them.20 A significant minority were first prepared to stand at least twice in lengthy queues at government offices to claim their ID books and thus the right to vote. A majority then stood in queues, first to register, then to vote. A society in which many people are prepared to comply with onerous official requirements to claim a civic right does not seem inherently ungovernable.

This does not mean that the problem of creating a sense of civic obligation is illusory. It is one thing to submit to public authority to claim a right which bestows a sense of efficacy and opportunity to express one’s identity; another to do the same for a more indirect public benefit with more obvious personal cost, such as paying electricity bills to ensure that the service is continued. But it may mean that the blocks of effective state-building are more available than many analyses assume.

Of course, the percentage poll reflects only those who registered. But the evidence suggests that the number of citizens willing to participate exceeded those able to do so. And again, that some two-thirds of the citizenry was willing to wend its way through a bureaucratic maze to vote suggests a significant ‘critical mass’ available for a state-building project.

That said, the fact that up to a third of citizens may not have voted and that the ANC’s 66% share of the vote may translate into the express support of little over 40% of adults, holds as many lessons as the democratic enthusiasm noted here.



If citizens are not moved to vote primarily by interest calculations, the tasks facing democratic government may be different to those assumed by many in South Africa’s elite.

The dominant view is an ideology of delivery which holds that, given the apartheid legacy of material inequality, citizens can only be induced to endorse democratic institutions by ‘delivery’ of goods and services. By implication, if democratic intangibles such as the right to vigorous representation must be compromised in the process, the gains in heightened citizen confidence will far outweigh any democratic losses. There are strong elements of instrumentality in this assumption; citizens are believed to see democracy as a source of material benefit, not self- expression.



The implied claim that there is a ‘trade-off’ between material improvement and democratic quality is dubious, given data from Southern countries indicating that democracy is not only compatible with growth and poverty reduction, but may be indispensable to both.21 But it also misreads South Africa’s citizenry and, therefore, its preconditions for democratic viability. A society in which most citizens are inclined to participate in democratic politics in part because they see it as an expression of their identity, is one in which the task of ‘winning society for democracy’ may be less onerous than implied, and one in which material delivery will not be enough.

The claim that South African political affiliations are shaped by identities is threatening to many politicians and intellectuals because it implies voter ‘irrationality’; the claim that the black majority is incapable of making rational political choices underpins racist ideology. This may explain why elite perceptions of the electorate constantly resort to instrumental criteria: racism’s opponents have a deep need to demonstrate that South African voters are ‘normal’ utility maximising citizens.



The response is understandable but flawed since it assumes that there is something ‘normal’ about the citizen who votes his or her interests rather than identity – that the calculations of rational choice theory are indeed a superior form of rationality.

But the utility maximising voter is itself an ideological creation. Not only do intangibles such as democratic commitment influence voter behaviour in all democracies: why else would people regularly vote for candidates or parties which have no hope of winning? But identities are important to choices in even the most seemingly instrumental democracy. In Britain, does the fact that some regions returned Labour candidates, others Tories, through a century mean that the majority make the same interest calculations every time or that their choice has something to do with their identity? And is being a US conservative or liberal purely an interest calculation or might it have something to do with who people think they are? And what of the historical importance of religion in voter choices in countries such as Holland? The Northern examples are selected to show that, even in democracies assumed to be ‘normal’ because elections are shaped by voters’ interest calculations, identity is important.

The importance of identity creates an opportunity for it means that the electorate is willing to see democracy, at least in part, as a ‘deliverer’ of intangibles such as self-expression rather than as purely a source of material benefit. In societies such as South Africa, that suggests a strategy for democratic strengthening very different to that which now dominates.

If most South African voters care as much about intangibles – of which identity is only part of a wider value placed on being heard – the challenge lies not in ‘delivering’ at the possible expense of self-expression, but in deepening and broadening the latter by strengthening participation and solidifying the relationship between legislators and citizens. It also, in the interests of economic growth and poverty reduction as well as democratic strengthening, requires a stress on difference as an asset rather than a liability, for only this can accommodate differing identities in a common political space.



Despite the electoral enthusiasm described here, the task of encouraging citizens into democracy remains important and this too is a powerful argument against one aspect of the ideology of delivery. A claimed message of the election is that citizens overwhelmingly conferred a mandate on the ANC to implement its agenda. Respect for difference may then appear unnecessary, since those who are different comprise at most a third of adults. But if those who explicitly endorsed the ANC at the polls are less than half the electorate, then most of the citizenry is still to endorse the government’s programme. This implies that extending the reach of the democratic state remains as, if not more, urgent than material ‘delivery’.

Democratic prospects will depend not primarily on improved public administration and enhanced ‘delivery’, but chiefly on the degree to which the necessary fight against poverty and inherited inequalities is pursued in ways which broaden and deepen the channels for democratic self-expression of all South Africa’s identities – in a manner which recognises that, even in post-apartheid South Africa, democratic intangibles matter to citizens as much as material improvements.




1. See Steven Friedman, Democracy, Inequality and the Reconstitution of Politics, Woodrow Wilson Center for International Scholars, Washington DC, 2000 (forthcoming).

2. S. Finkel, ‘The Effects of Participation on Political Efficacy and Political Support: Evidence from a West German Panel’, Journal of Politics 49 (2), 1987.

3. Steven Friedman and Louise Stack, ‘The Magic Moment’, in Steven Friedman and Doreen Atkinson (eds), The Small Miracle: South Africa’s Negotiated Settlement, Ravan Books, Johannesburg, 1995.

4. Claire Robertson, ‘Contesting the Contest: Negotiating the Electoral Machinery’, in Friedman and Atkinson, ibid.

5. Anthony Downs, An Economic Theory of Democracy, Harper and Row, New York, 1957.

6. Donald Green and Ian Shapiro, Pathologies of Rational Choice Theory, Yale University Press, New Haven, 1996.

7. Steven Friedman, ‘Agreeing to Differ: African Democracy – Its Obstacles and Pitfalls’, Social Research 66(3), Fall 1999.

8. Michael Bratton, ‘A First Look at Second Elections in Africa’, Transformation to a Successful Democracy, Institute for Federal Democracy, Konrad Adenauer Stiftung, Durban, December 1998.

9. Frances Fox Piven and Richard Cloward, Why Americans Don’t Vote, Pantheon, New York, 1988.

10. Graeme Gotz Buying, in Staying Out: The Politics of Registration for South Africa’s First Democratic Local Government Elections, Centre for Policy Studies, Johannesburg, October 1995.

11. Shaun Mackay, ‘IEC’s Sleight of Hand is Not in Electorate’s Long-Term Interest’, Synopsis, Centre for Policy Studies 3(1), March 1999.

12. The two most oft-cited polls found 83% expressing a voting intention. Rod Alence and Michael O’Donovan, If South Africa’s Second Democratic Election Had Been Held in March 1999: A Simulation of Participation and Party Support Patterns, mimeo, Human Sciences Research Council, Pretoria, 1999 , p. 7.

13. Centre for Policy Studies/National Business Initiative, Quarterly Trends, April 1998.

14. The exchange occurred at a seminar in Bonn, April 1998.

15. The theory in its ‘pure’ form insists that no one will vote unless they believe that their ballot alone will be decisive. While this perception was more likely in ‘close’ provinces, the difference in perceived utility, given that even in close elections it may not be rational to assume that one’s vote will be decisive, may not have been enough to influence behaviour.

16. Friedman and Stack, op cit.; Craig Charney, Voices of a New Democracy: African Expectations in the New South Africa, Centre for Policy Studies, Johannesburg, 1995.

17. See Rapport survey reported in Centre for Policy Studies/National Business Initiative, Quarterly Trends, April 1998.

18. SA Broadcasting Corporation, Election Special 2/6/99.

19. Understood in the sense it is used by Robert Putnam, Making Democracy Work: Civic Traditions in Modern Italy, Princeton University Press, Princeton, 1994, as an indicator of those attitudes of reciprocity and public education which are, in this view, at the core of a democratic culture.

20. This term is used by Michael Mann, States, War and Capitalism, Blackwell, Oxford, 1988, to describe the process by which the state brings society within its orbit.

21. Dani Rodrik, The New Global Economy and Developing Countries: Making Openness Work, Overseas Development Council, Washington DC, 1999.