On the high road to oblivion

Helen Suzman Foundation | Nov 30, 2001
Evidence from a recent South African opinion poll indicates that a leftwing breakaway from the ruling party would be doomed.
On the face of it South African electoral politics has become much more unpredictable and fluid than most observers would have expected a few months ago. But surface tensions can be misleading. The results of a national poll of public opinion conducted in July help us to grasp some of the implications of the split in the Democratic Alliance and the potential split within the ruling tripartite alliance.

Respondents were asked (among many other questions) which party they would support if a general election were to be held soon and what their second choice would be. In Figure 1 their responses are compared to those in the equivalent survey conducted just before the 1999 general election.

The results suggest that, between March 1999 and July this year, in the total adult population:
  •  the ANC lost a slight margin of support;
  • support for the DA — that is, the DP, NNP, FA and new DA combined —also dropped a little;
  • the Inkatha Freedom Party made significant gains;
  • most of the smaller parties lost support, with the Afrikaans ethnic parties all but obliterated;
  • the UDM, which had once looked potentially nationally viable, has been reduced to a rump of support in the Transkei;
  • the proportion of voters who said that they were uncertain about their party choice or had decided not to vote or even not to register rose by 5.4 per cent.
The significant rise in the latter group of people who were uncertain or did not plan to vote is partly due to the fact that, unlike in March 1999, there is no general election in the offing. But mid-term blues notwithstanding, the electorate is undoubtedly less politically engaged than it was two years ago. The experience of pollsters all over the world is that people who are uncommitted in pre-election polls are much less likely than those who make choices between parties to vote in elections. Those among the uncommitted who do eventually vote tend to distribute their votes in much the same way as the majority. As a result, if one excludes the uncommitted voters and recalculates the percentage distribution of choices among those who express party preferences, one has an approximation of what the support for parties would be likely to be if an election were held soon. Figure 2 shows what party support looks like after the responses to the March 1999 and July 2001 surveys have been recalculated leaving out the uncommitted. To illustrate that this operation does indeed approximate to voting outcomes, the March 1999 survey results are also compared to the actual election outcome in June 1999.

The slight differences between the recalculated March survey figures and the actual election results are explained by the parties’ relative turnout. Figure 2 shows us that:
  • the ANC performed slightly better in the June 1999 election, than the March survey indicated because a high proportion of its declared supporters went to the polls. In addition to high supporter commitment, elections on public holidays in areas in which there is very little else to do except to join the masses at the polling stations ensures this result;
 
  • the New National Party performed below its levels of declared support in 1999 mainly because a large proportion of its voters live in areas that are traditionally areas of exceptionally low voting turnout, such as the Cape Flats and the rural areas of the Western and Northern Cape provinces;
 
  •  the IFP performed slightly better in the actual elections than its levels of support in the survey — as it always has done. This is because very many of its voters are relatively apolitical, poor rural traditional people who are not effectively mobilised until just before election day itself;
 
  • The Freedom Front was in the process of losing support to the DP as the election campaign progressed. The 1999 election represented its point of final collapse.

If these voting patterns hold, the July 2001 poll suggests that in a hypothetical general election the ANC would maintain its position with just short of a two-thirds majority, while the DP/DA would remain secure in its position nationally as the major opposition party. Since then, of course, the alliance has been rocked by the acrimonious departure in October of NNP leader Marthinus van Schalkwyk and his top henchmen. In the July poll the extent of support for the DA’s separate components, the NNP and the DP, has to be interpreted from the proportions of voters who chose to name these parties rather than naming the DA as the overarching alliance. On this basis the New National Party has been virtually wiped out as a separate support base, dropping from some 8 per cent to barely more than 1 per cent. The DP on the other hand retains at least 50 per cent of its previous loyalty.

Looking at the second choices of respondents (not shown in the Figures), the DP is again likely to gain far more than it loses. The NNP supporters’ second choices were DP/DA: 47 per cent (39 per cent to the DP as such), uncertain: 11 per cent, UDM: 10 per cent, ANC: 5 per cent, PAC: 5 per cent, IFP 3 per cent, AEB: 2 per cent. Co-operating with the ANC after the split is therefore unlikely to bring the NNP any significant additional support.  Non-ANC voters appear to want strong and determined opposition and relatively few are interested in a party that will co-operate with the ANC from a position of weakness.

The situation is not much better for the NNP within its former stronghold among coloured people in the Western Cape. Here the NNP has declined to a separate support level some of 2.9 per cent, compared with 15.6 per cent for the ANC, 15.1 per cent for the DP and 25.9 per cent for the DA as the united entity. Even as a second choice in this constituency, the NNP attracts only 5.6 per cent; less than the DP (7 per cent) and the DA (9.5 per cent). With its separate identity so thoroughly eclipsed, it is very doubtful that the NNP is viable as an independent party or as a coalition partner for the ANC. The only outside hope for the NNP is that the circumstances of its break with the DA might have altered the image of the DP/DA to its advantage.

The split in the opposition alliance may have been public and messy, but is unlikely to alter the political landscape significantly. What of the strife between the ANC and its socialist Alliance partners, the South African Communist Party and the Congress of South African Trade Unions, which has been growing steadily for several years? By now one would have expected the left’s increasingly open criticism of the government over privatisation, Aids and Zimbabwe to have shown up in increased support for the SACP. Ever since 1994 a few respondents in MarkData polls have elected to mention support for the SACP rather than for the ANC. The level of this support has not changed over recent years and the results presented here show level of mention standing at a mere 0.3 per cent in both 1999 and 2001. Even the number of ANC supporters who opted for the SACP as a second choice in July was unmoved at a mere 6 per cent. Thus the notion that fundamental change in South African politics is likely to spring from within the ruling party itself is not apparent at the present time. If ever the SACP and Cosatu were to establish a separate progressive labour-based party, its fate would probably resemble that predicted here for the NNP.

Much interest is currently focused on the future outcome of an election in the Western Cape. Figure 3 shows party choice in the Western Cape in the July 2001 poll.

If the DP/DA loses no more than the people who specifically support the NNP, it is still likely to be the largest single party, however it would have to do a deal with the ACDP and the UDM to ensure a majority. If it loses some of the former NNP voters who appear as DA supporters in the results above, the ANC and the NNP together might emerge as the largest party, but they too would have to do a deal with one or another smaller party (the PAC, ACDP or UDM) to achieve a majority. The outcome is likely to depend on how effectively the various parties motivate their supporters to vote on the day. The only thing that appears certain at this stage is that the result will be a cliff-hanger.

In KwaZulu-Natal the IFP seems likely to translate its recent gains into a clear lead over the ANC in the province. These gains may partly be a response to the strengthening relationship between IFP leader Chief Mangosuthu Buthelezi and the Zulu monarch, King Goodwill Zwelithini, who recently reconfirmed Chief Buthelezi’s status as the traditional prime minister of the Zulu nation. Figure 4 shows the breakdown of support among voters in KwaZulu-Natal who made a choice in the July 2001 poll. The IFP is likely to govern the province in coalition either with the ANC, as at present, or the DA/DP. The choice of the DP/DA is unlikely unless Chief Buthelezi decides to relinquish his position as minister of home affairs in the national government and chooses instead to become premier of KwaZulu-Natal.    

The only other province in which the outcome might be in doubt is Gauteng. Figure 5 shows party choices in July. If the DP/DA persuades most of its supporters to go to the polls and the ANC has a poor turnout, then the result in a future election is likely to be in balance. The ANC might have to form a coalition with the either the IFP or the NNP to govern the province. The DP/DA has no obvious coalition partner unless it is achieves some kind of agreement with the IFP, which seems unlikely at this stage but cannot be ruled out.

The breakdown of voter choice in contested provinces illustrates how important relative turnout is becoming in our elections. Apathy and enthusiasm are by no means equally distributed at present. A MarkData poll following the 1999 election found that at least 14 per cent of adult South Africans interviewed had not registered to vote or had registered but did not vote. They broke down by race as follows:


            Africans:                       10 per cent

            Coloured                       30 per cent

            Indian                           32 per cent

            White (Afrikaans)           30 per cent

            White (English)             10 per cent

Those figures suggest that opposition parties attracting most of their support from non-African minorities were at a huge disadvantage due to the apathy factor. The major reasons whites gave for not voting were lack of interest in politics and the feeling that “voting is futile”. If these tendencies can be countered then the opposition will find itself with a quite considerable bonus of votes. A larger turnout among minorities in the key economic region of Gauteng, for example, could swing the province in favour of an opposition coalition. In addition, although 43 per cent of coloured and 38 per cent of Indian voters (with uncertain and non-voters excluded) currently support the ANC, their second choices in the July 2001 poll strongly favour the DA or DP — hence the DA could still expand quite significantly among non-African voters.

The DP/DA will remain as the major opposition party and it has some scope for expansion up to a ceiling of around 23 per cent without more African support. With determined effort it could retain control of the Western Cape and most of the local authorities that it now controls. To challenge the ANC nationally, however, would require it to increase its level of support among African voters. In the March 1999 polls the DP attracted less than 1 per cent of African support, by July 2001 this had increased to 2 per cent for the DP and DA combined (excluding 1 per cent support for the NNP among Africans). The second choices of African voters reflect an increase in support for the DA/DP from 4 per cent to 6 per cent. Hence there may be some progress but it will be painfully slow.

One could argue that these results show that the main feature of our electoral politics is its reassuring stability. After each episode of controversy, the same shape re-emerges. As far as the viability of multiparty democracy is concerned, however, such stability is about as comforting as a coffin at rest. The ANC’s very large and apparently enduring majority is deeply corrosive of the character and accountability of its leaders, while producing alienation and apathy among citizens who oppose it.

Floating voters are the lifeblood of any multiparty democracy, for these are the people who make change of government possible. In South Africa the only floating voters are coloureds and Indians, who are too few in number to ensure alternation in power. Only some 10 per cent of Africans entertain, as either a first or second choice, voting for a party with a “non-African” identity. Only about the same proportion of whites entertain supporting any party with an “African” identity. Thus the ANC, DP/DA and IFP need have no worry about losing core support to each other. It is in this feature of politics, above all, that the rot in our democracy will grow.