Ramaphosa and the Strange Workings of ANC Democracy

 R W JOHNSON

INTRODUCTION

It is difficult to imagine that there is any country in which people are more leadership-focused than South Africa. Very often one finds that voters identify their party by its leader, not just as a fact about that party but as signifying its very identity: Buthelezi was Inkatha; Mandela was the ANC; Tony Leon was the DP/ DA. One also notes that there is a great hunger to find the Providential Leader, a Smuts or a Mandela to whom one effectively surrenders one’s powers of judgement. Both white and black political traditions feed into this. The dominant African tradition is that of chiefly authority, while the trekboer tradition led to a similar veneration of strong leaders by whites.

One result of this political culture is an ever-present eagerness to believe that this or that figure – usually a new figure – will provide salvation. Currently the providential man is Cyril Ramaphosa. Not only do his ANC followers express extravagant hopes in him, but so too do any number of white journalists, businessmen and DA matrons. This is having real effects, as the ascent of the Rand shows. Ramaphosa’s embrace of “expropriation without compensation” – indeed, his declaration that this could produce “a garden of Eden” in South Africa – suggest that he shares this unrealistic euphoria. In fact, the first time that any farm is expropriated without compensation, all agricultural land in the country becomes valueless and unsaleable. Farmers will no longer have the collateral for their loans and most would go bust immediately, a huge consequent blow to the banks, to exports and to food security. In the ensuing chaos and famine, it is most unlikely that any government would survive. With “expropriation without compensation” the ANC has found its own form of Mutually Assured Destruction and Ramaphosa, like Dr Strangelove, seems to have learned to love the Bomb.

The ANC leadership contest has shed a strong light on the current state of the movement. In the months before the conference, the eNCA TV station used the services of Markdata to conduct a large opinion survey among some 5 000 people, including several thousand ANC voters. Focus groups were also convened in various parts of the country. The results were striking. ANC voters everywhere said that President Jacob Zuma was a disgrace: he had pulled the country down and brought pervasive corruption, the cause of every woe. He had to be got rid of – and then the ANC would self-correct and again become the ANC of Mandela. So they would vote ANC again: after all, how could they vote for the opposition Democratic Alliance which (ridiculously) they thought would abolish all welfare payments and restore apartheid? The results of this survey are published in the sections which follow, for they provide an unrivalled picture of the country’s mood on the cusp of a great change.

When ANC voters were asked who they would prefer as their new president, 48.4% said Ramaphosa and 21% Zuma’s ex-wife, Dr Nkosazana Dlamini-Zuma. Of the nine provinces Ramaphosa led in eight, the exception being Zuma’s bastion, KwaZulu-Natal. When one looked at the tribal breakdown, Ramaphosa led by 3:1 among the Sotho, by 5:1 among the Tswana, by 9:1 among the Pedi, and by 8:1 among minority tribes. Among Xhosas he led by better than 2:1 and even among Zulus his support stood at 32.8%, as opposed to 37.5% for Dlamini-Zuma. That was the situation as of September 2017, but when the pollsters returned to some of their sample in late November, the Ramaphosa momentum had continued and he now led Dlamini-Zuma by 64% to 14%.

The pollsters also asked ANC voters whether they preferred the “radical economic transformation” policies of Zuma/Dlamini-Zuma, aimed at a complete redistribution of all wealth and income, or whether they would prefer a more pro-business orientation in the hope that that would bring more jobs. Overall, African voters preferred the latter by a 5:2 majority. Although Ramaphosa’s image was that of a wealthy businessman supported by other businessmen, the pollsters found that he was supported by large majorities among the less educated, the poorer and the unemployed. The reason is obvious: under ANC rule the number of unemployed people has risen from 3.7 million to 9.4 million. Inevitably, jobs are by far the No.1 issue – and Ramaphosa was thought more likely to mean more jobs.

Then came the ANC conference where one saw the full power of the regional patronage barons, of Zuma’s incumbency, and of the bribery and manipulation of the delegates from the ANC branches. As a result Ramaphosa won, by the narrowest of margins, what should have been a landslide. The result turned on three provinces in particular – Mpumalanga, the Free State and the North West. The pollsters had found that Ramaphosa led in all three by 3:1 or 4:1, but the Free State and the North West delivered their bloc votes to Dlamini-Zuma and Mpumalanga dithered, at first supporting a (fictional) Unity candidate and then sliding onto Ramaphosa’s side at the last moment. This last switch was what decided the contest and it left Zuma visibly thunderstruck and unmanned. His daughter, who had been sitting elsewhere in the hall, quickly went and sat next to her father, knowing that his hopes had been dashed. David Mabuza, the premier of Mpumalanga, had clearly made a deal with the Zuma camp and then got a better offer from Ramaphosa.

In addition, the conference voted for radical economic transformation including the expropriation of land without compensation – exactly the policy which bankrupted Zimbabwe.

One cannot understand the results without realising that power in today’s ANC lies largely with great regional barons who control patronage, jobs, tenders and contracts. Three such barons are particularly visible: David Mabuza of Mpumalanga, Ace Magashule of the Free State and Supra Mahumapelo of the North West. All three men rule their provinces with a rod of iron. In both the North West and Mpumalanga, what bribery and patronage can’t do is sometimes settled by assassination. In some cases the liberation “war veterans” – as in Zimbabwe, actually young thugs drilled as a private army – bullied and intimidated those who did not take the Zuma side. Magashule is deeply involved with the Gupta family, who have been Zuma’s main sponsors and beneficiaries. Thanks to the manipulation of conference votes, he has now become the ANC’s Secretary-General – the boss in charge of the entire organisation.

There were no angels here. To win Ramaphosa had to fight fire with fire. In the run-up to the conference, he raised hundreds of millions of Rands from white business (leading the Zuma forces to depict him as “the candidate of the imperialists, of white monopoly capital”). One has to ask what this money was used for. After all, Ramaphosa is worth at least $420 million himself. The campaign was simply a matter of public appearances and speeches. There were no leaflets, no posters, no TV ads. So what could all that money be spent on other than financial inducements of one sort or another? It is a very safe bet, in other words, that not only David Mabuza but many ANC delegates went home from the conference a whole lot richer. For most of them after all, this would have been the primary reason to join the ANC.

It is a peculiar thing to watch a political party nominate by a hair’s breadth a candidate preferred by its voters by a huge majority and then endorse economic policies which are precisely the opposite of what its own party voters want. Is it really the case that the activists who make conference policy are so different from ANC voters at large? It seems unlikely.

The Zuma strategy lay in the slogan of “radical economic transformation” and the campaign against “white monopoly capital”. His last-minute announcement of free higher education was clearly supposed to show evidence of what radical economic transformation might look like, viz. a purely voluntarist radicalism not based on budgeting or rational economic calculation. The conference vote for the expropriation of land without compensation showed that this approach to policy-making is far from dead in the ANC. But what is interesting about this strategy is the assumption that the way to mobilise and win over conference delegates is via appeals for racial solidarity against whites (all landowners being presumed to be white). Thabo Mbeki, it should be remembered, made the same assumption: he abandoned the non-racial rhetoric of the Mandela period for continuous appeals for black solidarity against whites. The powerful thing about this sort of racial populism is only partly due to black hostility to whites (which opinion surveys suggest is far from universal) and much more to the implication that any black who does not rally to this cry is proving himor herself to be a willing pawn of the whites and thus standing outside and against other blacks. Most blacks are extremely nervous of such an accusation, not only because African tradition emphasises the complete consensus of the village community but because no one has forgotten the fate meted out during the Struggle to those suspected of such disloyalty: within ANC circles the most dread accusation was of “selling out”.

The most striking manipulation was seen with the bloc votes from the Free State, the North West and Mpumalanga. The Free State’s vote was decided at an almost certainly illegal meeting of the provincial general council held with not much more than one day’s notice. The premier, Ace Magashule, is not only openly in league with the Guptas but has often distributed to his ANC followers largesse deriving from the Guptas. In the North West, opponents of the premier, Supra Mahumapelo, have been driven out and, in some cases, gunned down. More than forty ANC branches that supported Ramaphosa were simply dissolved and replaced by fake new branches of Mahumapelo’s puppets. In Mpumalanga too, opponents of the premier, David Mabuza, have been known to meet a violent end: journalists joke uneasily that Mabuza “runs the most efficient police state in Africa”.

This is the violent reality of power in three of South Africa’s nine provinces, but their violence is easily surpassed by KwaZulu-Natal where scores of ANC officials and activists have been murdered. Things are relatively better in the three Cape provinces (West, East and North), but it is a matter of degree rather than of kind. One cannot but be reminded of the scenario painted by IRR CEO Frans Cronje, who depicts a future in which the black and white middle classes continue to live in a privileged bubble in the big cities, while the rest of the country is ruled either by traditional chiefs or criminal gangs. We are already some way towards that.

One of the interesting features of the focus groups conducted by Markdata was the nostalgia for the old Bantustans that they uncovered in several places. The reasons given were twofold: first, the Bantustans had brought many jobs – a local civil service, police force, broadcasting authority and airline, for example – almost all of which had vanished after 1994; and second, there was the possibility of access to power within that Bantustan, whereas after 1994 all power was transferred to (inaccessible) Pretoria.

But as one surveys the contemporary landscape one can see another reason. The Bantustan rulers in the North West, the Free State and Mpumalanga were all much less oppressive than their present-day replacements, while Buthelezi’s rule in the KwaZulu “homeland” was not accompanied by anything like the degree of violence which marks contemporary KwaZulu-Natal. Similarly, today’s large black townships are all far less controlled by the police than they were thirty or forty years ago. Today’s police are far more corrupt, ill-disciplined and simply absent, leaving a much freer field for the gangs. This, then, is the uncomfortable reality. No one believes that the old apartheid order was fair or just, but the hope was that it would be superseded by something better. As one looks at the lived reality of black life in South Africa’s townships and rural areas today – the gangs, the oppressive local bosses, the violence, the mountainous unemployment and the often tyrannical chiefs – it is difficult to believe that this has happened.

Note on methodology

The opinion survey was conducted by Markdata (Pty) Ltd, an organisation with some 30 years’ experience in conducting field surveys for public, private, and civil society organisations. It was carried out as part of an ‘omnibus’ survey and at the behest of eNCA.

The survey was conducted in September 2017 among a fully representative national sample of just under 5 000 people. This sample was carefully set against the latest census data for each region in order to achieve full representivity. Interviews were conducted in person (none of them by phone), and in rural as well as urban areas. All respondents were interviewed in their first language by experienced interviewers from the same language group.

In addition, focus groups were organised in the Eastern Cape, KwaZulu-Natal, Mpumalanga, and Gauteng. In each of these provinces, four focus groups were convened. This was done in order to amplify the data from the survey because one always learns other things from conversations in depth. However, the results cited in the national and provincial reports which follow are taken solely from the survey data, and not from the additional insights provided by these focus groups.

As noted in our national report, (see Survey results across the nation), while we were conducting our major survey in September 2017, we counted 2 717 respondents who were ANC voters. We asked them whether or not they would be willing to act as a panel to which we could return at a later stage with a few extra questions. Our aim here was to test the evolution of opinion, and also how to see how these voters were responding to new issues as they arose.

This request produced a panel of 1 198 ANC voters. It is important to say that, because this panel was self-selected, it was likely not to be quite as perfectly representative as our overall sample. Among other things, supporters of Cyril Ramaphosa might have been disproportionately likely to volunteer to participate. We returned to this panel late in November 2017 and asked them various further questions, as set out in the national report which follows.

SURVEY RESULTS ACROSS

 THE NATION

Based on the survey results obtained in the nine provinces (as set out in the sections that follow), it is possible to stand back and look at the national picture as a whole. We started off by asking respondents how they had voted in the last parliamentary and local elections, which produced the following breakdowns by race:

Voting in last two elections

As expected, this showed a picture of considerable racial polarisation, although it should be noted that the African figure for abstention is far too low – and the Asian figure far too high. In practice we know that some 25% to 30% of Africans did not vote: the lower figure here results from the fact that many Africans treat this question simply as a matter of partisanship or even of identity – e.g. “I am ANC”. Asians, on the other hand, are notoriously cagey and are the most prone either to insist that their vote was secret or that they did not vote.

We then looked at the same figures by gender and by age:-

The relationships here are less well known. The net ANC advantage among men is 28.2%, but among women this swells to 35.2%. This feminine “bulge” in ANC support is highly significant, particularly as women constitute 51.9% of the electorate.

It is no surprise to see the high abstention rate amongst the youngest category of voters, but the striking thing is that this leaves the ANC with less than a 2:1 advantage in this group – whereas in the oldest age group (and, in fact, in all other age groups) the ANC advantage is 5:2. Again, this is a very significant difference when one considers how biased towards young people the South African electorate is. One of the biggest questions for the future is how these young voters will behave as they get older: will they continue to abstain, fall into the ANC-voting pattern of their elders or defect to other parties?

A good pointer to the mood of the different racial groups is provided by our question about how they felt race relations had developed in the recent period:

As can be seen, African voters feel unaffected by the campaign against “white monopoly capital” or the general search to expose white racism of any kind. The result is that they are far more optimistic about the state of race relations than any other group. The high number of Asians saying race relations are “much worse” is a key signal that Asians are by some margin the most alienated group. By comparison, 39% of whites thought race relations were much or a bit better and 40.4% said they were much or a bit worse, which is an almost completely even split. This contrasts with 48.1% of Asians saying race relations are worse and only 33.2% saying they are better.

We then broke down the choices for ANC president by race, but noted these only where preferences reached 2.5% or more.

It is noticeable that large numbers of all the racial minorities expressed no opinion or said “Don’t know”, clearly feeling that choosing an ANC president lay outside their sphere of interest. As may be seen, Ramaphosa is not only by a large margin the most popular candidate but the only one with a fair measure of support across all racial groups. In fact, if one looks solely at the opinions of voters, including ANC voters, Ramaphosa’s lead is so clear that it seems surprising that the issue was ever in doubt.

One must realise, too, that the tradition that the ANC deputy president should succeed to the ANC presidency is very strong: indeed, many think it was Thabo Mbeki’s attempt to overturn that tradition that led to his being deposed by Jacob Zuma. Zamani Saul, the new ANC leader in the Northern Cape, spoke for many when he insisted that this tradition, which has been observed throughout the last fifty years, is the key to party unity and must accordingly be preserved. Saul, who was pro-Zuma in 2007 at Polokwane, is pro-Ramaphosa now. He even cites a resolution of the 1949 ANC conference which declared that “the Deputy-President shall preferably succeed an outgoing President.”

The fact that several women candidates for the ANC presidency made much of the gender issue led us to also examine the figures by gender and age:

As may be seen, the three female candidates all fared better among women. This was particularly so for Dlamini-Zuma, whose support was also much the same across both age groups. This contrasts with Ramaphosa, whose support was notably more masculine and older – even though he led comfortably in all categories. It is tempting to connect this to our experience in some of our focus groups, where older men often found it difficult to support a feminine candidate. It is worth noting that a quarter of all ANC voters were unwilling or unable to express a preference for any candidate. This rather suggests that the higher feminine score for Mbete, Sisulu and Dlamini-Zuma was not necessarily the reason for Ramaphosa’s lower level among women, for there was still a plentiful pool of women not opting for female (or any other) candidates.

Inevitably, there has been a good deal of talk about tribalism and, particularly, “the Zulu factor”, so we thought it would be worth looking at how the different African language groups (ie. first language-speakers) divided as between the two top candidates, looking only at presidential first choices. The percentages in the table below are the proportions of each language group preferring Ramaphosa or Dlamini-Zuma:

This is nothing like the acute tribal polarisation which was so often depicted in the run up to the Nasrec conference. Dlamini-Zuma is ahead among isiZulu-speakers, but not by much. She gets some useful votes from isiXhosa and Sesotho-speakers, but her problem is that she leads only within her own language group and lags badly in all others. Ramaphosa impresses again by the breadth of his support – just as he was the only candidate with significant support across all racial groups, so he is the only one with substantial support from all African language groups. The frequent suggestion that a Ramaphosa victory would lead to serious trouble among isiZulu-speakers or in KwaZulu-Natal, looks unfounded.

On the other hand, Zulu opinion is notably more charitable towards Zuma himself. While 46.1% of isiZulu-speakers thought there was a corrupt relationship between Zuma and the Guptas, a third (33.2%) did not – a much more favourable view than any found among other groups. And when asked what should happen if a corrupt relationship were proven, only 27.5% of isiZulu-speakers thought both Zuma and the Guptas should go to jail. Far more (39.4%) wanted just the Guptas punished, though 22.5% said “Don’t Know”. While these opinions are far more favourable to Zuma than those found among other groups, Zulu opinion is fragmented and far from solid behind the President.

We then explored preferences for ANC president along different dimensions. In terms of image, the two principal candidates were in strong contrast. Dlamini-Zuma stands for more radical, more clearly socialist policies, while Ramaphosa has a more moderate, business-friendly image. This often led to an assumption that Dlamini-Zuma’s support would be biased towards poorer and less educated ANC voters, and Ramaphosa’s towards better educated and more middle class ones. In fact, the exact opposite is the truth. Here, we divided our ANC-voting sample into two groups: those with no, primary or only some secondary education, and those with matrics and higher education:

Thus while Ramaphosa has a clear advantage across both groups, his lead is far bigger among the less educated. We then checked these results by comparing the presidential preferences of the ANC-voting unemployed with those employed in the formal sector.

Finally, we separated out those ANC voters earning between R0 and R4 999 a month, and those earning R10 000 a month and over and examined their presidential preferences.

Again, Ramaphosa has a clear advantage across all groups, but Dlamini-Zuma scores notably better among the better educated, the formal sector employed and the higher earners. It is often argued that the better educated are more prone to take up ideological positions – and there may be something in this. However, we would suggest a much more basic explanation. Unemployment, which has more than doubled under ANC rule, is overwhelmingly the chief cause of poverty and is by far the leading political issue. Thus far, for all its pro-poor rhetoric, the ANC has presided over increasing poverty and unemployment – and almost all the advantages of ANC rule have gone to the better educated and better off, who are able to benefit from higher public sector employment, affirmative action, and BEE. This is the chief cause of growing inequality in South Africa. As we have seen, there is a large majority among ANC voters favouring a more pro-business stance by the party in the belief that this will create more jobs, and Ramaphosa benefits from this. Crudely put, his more moderate position is seen as best for jobs – and this is decisive.

This led us to look again at voter preferences on whether the ANC should adopt either a more radical or a more pro-business stance in future. We found no differences across age groups, but it is worth showing gender differences and the picture that emerges among Africans – the group with the highest propensity to favour more radical solutions:

These are striking figures. Only one tenth of the electorate opts for the policy mix as now – an utterly overwhelming sign of dissatisfaction with the status quo among all age, racial and gender groups. Secondly, even among African voters there is a heavy 5:2 majority against more radical policies – which are actively desired by less than one African in five. Among women, the preference for more radical policies is particularly low: men preferred pro-business policies by a margin of 29.1%, but among women that figure rose to 35.2%.

These figures highlight a central conundrum in the current political situation. Both President Zuma and Dlamini-Zuma espoused the campaign against “white monopoly capital” and advocated “radical economic transformation” on the assumption that these “revolutionary” appeals would win the hearts of ANC activists and conference delegates. A subset of ANC activists may also be more radical than the generality of voters (or just ANC voters) we looked at. It seems peculiar to assume, however, that the way to court popularity among ANC activists is to espouse policies that lie flatly against the preferences of the ANC electorate as a whole. It looks, in fact, as if much of the ANC leadership has completely lost contact with what most ANC voters think and may not even be conscious of the huge divide that separates their assumptions from those of their electorate. It also seems unlikely in a democracy that such a situation is sustainable.

There also seems to be a similar disconnect over the choice of the next ANC president. In the run-up to the conference, the press was full of estimates that Ramaphosa and Dlamini-Zuma were “50-50 opponents”, “neck and neck” and that the race was “too close to call”. These estimates varied remarkably from what ordinary voters of all races wanted, including ANC voters, for our survey showed that among ordinary ANC voters Ramaphosa was well ahead in eight out of nine provinces. Even in the ninth one, KwaZuluNatal, there was no decisive majority favouring Dlamini-Zuma. Ramaphosa was also the only candidate with a strong measure of support across all racial and language groups.

Against this background, the Free State provides a good example of how the ANC’s electoral process seems actually to work. At a surprise meeting of the provincial general council (PGC) – called at twelve hours’ notice – Ace Magashule pushed through the nomination of Dlamini-Zuma for ANC president, citing 209 branch nominations for her as against only 44 for Ramaphosa. Yet ANC voters in the Free State preferred Ramaphosa to Dlamini-Zuma by 48.9% to 19.9%. The leader of the Ramaphosa faction, Thabo Manyoni, stayed away from the PEC meeting in protest and a court order setting it aside was later obtained. As a result, 27 members of the PEC were barred from voting at Nasrec, as were various branch delegates.

This Free State result holds several lessons. One is that the views of ordinary voters – even ANC voters – do not necessarily have much impact on the results of ANC electoral conferences. Second, the Free State vote, as secured by Ace Magashule, shows how well the Zuma regime has suited the Premier League. Moreover, while the rest of the country emphatically wants a change from that regime, the Premier League wants quite the opposite.

Re-assessing the views of a panel of ANC voters

When we were conducting our initial major survey, we counted 2 717 respondents who were ANC voters. We asked them whether or not they would be willing to act as a panel to which we could return at a later stage with a few extra questions, with a view to testing the evolution of opinion and how these voters were responding to new issues which might have arisen. This produced a panel of 1 198 ANC voters. Because this panel was self-selected, it was likely to be less perfectly representative than our overall sample.

We returned to this panel in late November, and began by asking once again for their preferences in the presidential race. When we first asked this question in September, the overall position among our initial national sample of ANC voters (2 717) had been 48.4% for Ramaphosa, 21% for Dlamini-Zuma, 4.3% for Sisulu, and various fractions of less than 4% each for the other candidates, with 13.5% refusing to express a preference for any of the seven leading candidates.

But when we asked this question again of the smaller number of ANC voters in our self-selected panel (1 198), the picture had changed substantially. The number refusing to express a preference had fallen to 2.1% and it seemed that all these late-deciders had moved into the Ramaphosa camp, for Ramaphosa now had 64%, Dlamini-Zuma had fallen to 14% and Sisulu, in third place, had risen to 6.7%. It is possible, however, that Ramaphosa supporters might have been disproportionately likely to volunteer to be on our panel. There is simply no way of knowing that. The most likely interpretation is that there could well have been some sampling bias of this sort, but there must also have been a genuine movement of opinion towards Ramaphosa.

What seems to have happened is that, as the race entered the final straight, there was increasing recognition that this was a two-horse contest and that a genuine momentum had developed towards Ramaphosa. The fact that in these two months the number of Dlamini-Zuma supporters had actually fallen by one third is highly suggestive. This, of course, makes it all the more remarkable that the race was so equal when the votes were counted at the Nasrec conference. This suggests a quite heroic level of delegate manipulation.

In further probing the views of our panel, we began by pointing out that, whereas the traditional ANC system had meant a single name being presented for the party conference to ratify, this had now evolved into a far more competitive system with multiple candidates and open campaigning. We asked how far voters approved of the new system, or if they preferred the old system:

As may be seen, opinion is split fairly evenly, but it would seem that the new way of doing things is here to stay. Among ANC voters, 62.3% either prefer it or have no objection to it, while the youngest age group is notably more in favour of it than the oldest age group. This guarantees that opinion will continue to move in favour of the new system as this younger group replaces its elders.

The most striking finding, however, was that approval levels for the new system were substantially higher in the North West and Mpumalanga. This is surely not accidental, given that these provinces are home to the two most “repressive” ANC regimes. As we will see from the provincial analyses that follow, high “nonresponse” levels in Mpumalanga probably reflect high levels of intimidation, a suggestion reinforced by Mathews Phosa’s allegations that groups of armed men have been roaming that province bullying voters into line. Similar accusations have been made in the North West, where a number of deaths have been attributed to the work of hit squads. Given that background, it is worth noting that ANC voters there are the keenest of all on an open, competitive system.

We then took this question a stage further by asking the panel for their preferences on the various ways of electing an ANC president: the old, closed system with one nomination; the current system with provincial delegates authorised by an ANC Provincial General Council (PGC); a system in which every branch votes irrespective of province; and a system in which every ANC member has a vote:

Again, though opinion is fragmented, it is striking that no less than 44.5% of ANC voters would opt for a greater degree of direct democracy than exists now. It may well be that this is a trend which party leaders will find hard to resist. The broader the franchise becomes in the future, the more difficult it will be to manipulate the vote. It is also striking that, in Mpumalanga and the North West, well over half of ANC members would opt for direct democracy solutions – and this despite that fact that Mpumalanga premier David Mabuza effectively urged a return to the old closed system with his call for a single “unity” candidate for each post. This not only runs against the general evolution towards a more open, competitive system – less than 30% want to see a return to such a system – but has been overwhelmingly disavowed by Mpumalanga ANC voters, among whom only a record low 12.5% want to see a return to such a system.

We then asked respondents how they would react if, as was often suggested in the media, delegates to the ANC conference – through the use of bribery, corruption and other pressures – were manipulated into choosing a candidate who was not really the choice of ANC members. In that case, said 29.1%, they would nonetheless loyally support whoever had been nominated. However, it was already clear to most that Ramaphosa was the popular choice and that, if this were to happen, Dlamini-Zuma would be the beneficiary. This was clearly in the minds of ANC voters in Limpopo (Ramaphosa’s home province), where only 17.1% said they would still be loyal to whoever was chosen, and even more to ANC voters in Mpumalanga, who had seen their preferences set aside by their premier, David Mabuza: here only 13.5% said they would remain loyal in any event. The old, ultra-disciplined ANC is clearly a thing of the past. Younger voters are noticeably less likely to stay loyal in such circumstances case than are older voters.

In addition, 36.1% of all ANC voters said that, in the case of such a manipulated result, ANC branches should protest and demand the holding of a properly free and fair election. A further 14.2% said that such an event might result in a party split, even though they themselves would not support that. The numbers mentioning the possibility of such a split rose to nearly a quarter of all respondents in Limpopo and Mpumalanga. There was also a 16.6% group of ANC voters who said that, in such an event, there might be a split which they would support if their own candidate had been unjustly treated. The number saying they might support a split in these circumstances rose to 26.4% in Mpumalanga.

The striking thing about these figures is that the old knee-jerk party discipline is now largely gone, while the bonds of solidarity have been greatly loosened. In effect, David Mabuza, in urging a unity ticket and single-candidate contests, was calling for a return to the past. However, this was contested throughout the party and particularly in his own province, where there is visible dissatisfaction with his “party boss” style. This suggests that a different vision of the ANC has grown up after more than a quarter century of open political life. Party voters see no need for the disciplined habits of clandestinity any more, and they want the party to live up to its claims of internal democracy. Typically, they look to the Mandela period as having exemplified the sort of ANC they want and to which they wish to return.

Just as we were talking to our panel, the events leading to President Robert Mugabe’s resignation unfolded in Zimbabwe. We thus pointed out that it was Mugabe’s attempt to promote his wife Grace as his successor which had proved to be the last straw. How far, we asked, did the panel feel that there was any comparison with the way President Zuma was attempting to promote his ex-wife to be his successor? If he achieved this, would there be trouble here? Extraordinarily, 12.9% of ANC voters said that, if Zuma managed to do this, it would simply be accepted by the party. However, another 35.9% simply dismissed the comparison, saying the situation here was different. But the biggest bloc of all, 45.2%, thought that if this occurred there would indeed be trouble.

This figure included over half (52.8%) of older (55+ years) ANC voters. The figure rose even further, to 55.3% in the North West, where premier Supra Mahumapelo has ridden rough-shod over the preferences of local ANC voters. It rose even higher, to 60%, in Limpopo, Ramaphosa’s home province, and was higher still, at 62.5%, in Mpumalanga – where David Mabuza’s bossism has generated much resentment. Strikingly, this figure sank to only 30.5% in KwaZulu-Natal, Dlamini-Zuma’s home province.

Finally, we tried to gauge the impact on the ANC’s electoral support if Dlamini-Zuma were to be pushed in. Though we refrained from telling the panel that a large majority of ANC voters clearly preferred Ramaphosa, more than half (52.4%) of the panel’s members nevertheless said that such an outcome would result in the ANC losing votes. Only 21.3% said that such an outcome would have no influence on the ANC’s electoral fortunes. KwaZulu-Natal was again a special case – only 28.5% of ANC voters there thought that such an outcome would lead to lost votes. That figure also rose to 63.3% in the North West, to 74.5% in the Northern Cape, and to 81.1% in the Eastern Cape. The very high figures in the latter two provinces reflected, of course, their strong pro-Ramaphosa leanings.

These figures are so suggestive in their implications that one has to marvel at the careless way in which President Zuma and his faction have been steering their party towards the rocks. By ignoring, indeed flouting, the clear majority opinions of ANC voters, the Zuma faction has risked huge damage to the ANC. What explains this?

One must remember that, for most of its existence, the ANC was not part of a competitive electoral system. When the party first embarked on an electoral role, it was still buoyed up by the excitement of liberation, the huge satisfaction of victory and the pleasures of power. For many years it felt that this was enough, that there was no need for any change to the way in which the party did things. But bit by bit the effects of being part of an open democratic system have fed through.

The old party leadership has clung to its old assumptions – that the ANC will “rule till Jesus comes” – because its voting bloc is so solid and so impervious to short-term trends. The municipal elections of 2016 showed, however, that those days are fading. Despite that, the Zuma faction has still been operating on the assumption that the ANC will simply accept whichever leader is imposed on it – and that this will be without electoral cost. These assumptions seem increasingly fallible.

SURVEY RESULTS IN KWAZULU-NATAL

Our first provincial report focused on KwaZulu-Natal (KZN), which was a key province in every sense. It was likely to have more delegates at the Nasrec conference than any other province, and it had long been the cornerstone of President Zuma’s support. But there was also a pro-Ramaphosa team in KZN, which was centred around former premier Senzo Mchunu. Mchunu had been evicted from office in May 2016 so that a new team could be installed under a new premier, Willies Mchunu, a strong supporter of DlaminiZuma. Towards the end of 2017, however, the 2015 provincial ANC conference – which had created the foundation for such a coup – was invalidated by the Pietermaritzburg high court. Senzo Mchunu’s team was confident it would win at any re-staged conference but, if this failed to happen, it was also possible that some KZN delegates might be disqualified from voting (as in fact later transpired when 27 members of the provincial leadership were barred from voting at Nasrec). The outcomes of the conference thus seemed to hinge largely on what happened in KwaZulu-Natal. Accordingly, we concentrated our attention on ANC (and other) voters here.

We began by asking which voters had voted for the ANC in the last national and local elections. The results re-affirmed that KZN is a heavily ANC province: 57.9% had voted ANC, while 18.8% refused to answer or said they had abstained from voting in at least one of these elections. By contrast, 22.8% had voted for Opposition parties. We then concentrated our attention on declared ANC voters, as the KZN delegation to the Nasrec conference would most likely be chosen from this group.

It was clear from our focus groups that voters saw corruption as the primary cause of South Africa’s economic woes. So we first asked voters how well the KZN provincial government had dealt with corruption. Among ANC voters, 41.8% said they had dealt quite or very well with the issue, while 56.3% said they had dealt not very well or not at all with the problem. The figure of 41.8% was quite high in comparison with what we found in other provinces. It was the first sign of a large bloc of ANC voters in KZN (40% to 45%) who clearly took a pro-Zuma and pro-Dlamini-Zuma line on most questions. These figures contrast strongly with the views of Opposition voters: only 27.1% thought the provincial government had dealt quite or very well with the issue, while 70.9% said they had dealt not very well or not at all with it.

When we asked how well the province’s government had dealt with poverty, the pattern was the same with 42.9% of ANC voters saying they had dealt very or quite well with the issue, and 55.8% saying they had dealt not very well or not at all with the matter. Among Opposition voters the figures were 29.7% and 68.5% respectively. Again, the 42.9% figure is rather high compared to other provinces.

When we asked the same question regarding land reform, again 40.6% of ANC voters said the province’s government had dealt very or quite well with the issue, while 50.2% said they had dealt not very well or not at all with it. Among Opposition voters, the figures were 32.1% and 65.6% respectively.

However, when we asked how well the provincial government had dealt with factionalism, the results were very different: only 28.7% of ANC voters said they had dealt quite or very well with the issue, which was a lower figure than the 30.9% of Opposition voters who gave such a reply. By contrast, a massive 60.3 % of ANC voters (and 57.7% of Opposition voters) said they had dealt not very well or not at all with the problem. This was almost certainly a reflection of the bitter factional struggles still going on within the ANC in the province.

The picture was much the same when we asked voters how well their province’s government had dealt with political violence. Only 24% of ANC voters (and 30.9% of Opposition voters) said they had dealt very or quite well with the issue, compared to 70% who said they had dealt not very well or not at all with the matter.

Overall, even ANC voters gave their province’s government a negative report card. However, there was also still a large (and probably well drilled and mobilised) core of voters supporting the pro-Zuma team still in charge of the province.

We then asked voters to rate their premier, Willies Mchunu. An astonishing 49.5% of ANC voters said he was doing a good job and deserved to continue – in most provinces, premiers had a mere 20% or fewer saying this. Again, this is probably a sign of a well-mobilised core of pro-Zuma voters. Our focus groups also showed that most voters were far more aware of their mayors and councillors than they were of their premiers, so it is difficult to believe that half of ANC voters were well acquainted with Willies Mchunu’s merits. Opposition voters were spread all over the place – 11.3% said “Don’t know” or refused to answer, 14% said Mchunu was not doing a good job and should go, while 25.2% said he was doing a reasonable job but they’d still like a change.

Against this background – and given the importance of corruption as an issue of key concern to voters – we then asked voters how hard they thought their provincial government had fought corruption. Clearly the politically correct answer for those who had just shown such support for Willies Mchunu was to say that it had fought it very hard – but only 17.1% of ANC voters said that. Another 26.8% (the rest of the pro-Mchunu bloc) said that the province’s government had fought corruption a bit but not enough. A further 20.6% of ANC voters said the government had made “very little effort” to fight corruption. But the biggest bloc (30.7% of ANC voters) said that either that the provincial government had made no effort at all or was corrupt itself. Among Opposition voters this last bloc rose to 45.1%.

Corruption was thus seen as a fundamental issue across the board. At least half of even the province’s ANC voters were very upset about it and anger over the issue was strong enough to fragment the otherwise unified pro-Zuma bloc. Outside the ranks of the ANC, the resentment against corruption was even stronger.

We then asked voters to give us their first and second choices of the candidates for the ANC president. At that stage, the only clear candidates were Dlamini-Zuma, Ramaphosa, Mathews Phosa, Jeff Radebe, Lindiwe Sisulu and Baleka Mbete, but we added the name of Dr Zweli Mkhize. However, as our focus groups had showed, there was considerable uncertainty as to whether Mkhize was really a candidate and this reduced his support.

Four significant blocs emerged among ANC voters:

(1) No less than 44.3% made Dlamini-Zuma their first choice – this being the same reliable pro-Zuma bloc. Apart from that she had little support, for only 10.3% made her their second choice.

(2) By contrast, 24.3% made Ramaphosa their first choice, while another 16% made him their second choice.

(3) Some 14.1% said they favoured no candidate. The large size of this bloc probably reflects the fierceness of factional conflict in the province and a corresponding desire to stay out of it.

(4) Roughly 8.6% made Zweli Mkhize their first choice and more than double that number made him their second choice, reflecting a considerable penumbra of potential support.

If we look further at the roughly 45% pro-Zuma bloc revealed by this data, it is almost entirely African. A full half of all ANC women support this group, as opposed to 30% of men. This bloc is also heavily reliant on older voters, with the over 45s making up no less than 44.6% of the bloc. The pro-Zuma forces have thus successfully mobilised older, Zulu-speaking and less educated women, particularly in rural areas. This is exactly the group which previously made up a large share of the IFP’s support in KZN. The big difference, of course, is that Durban is the last great ANC urban bastion (and patronage honeypot) and thus contributes many activists to the Zuma cause, an advantage which the IFP never had.

KZN has a heavily feminine population (59.3% are women), but gender was not a major determinant. Some 35% of women supported Dlamini-Zuma (a local woman, after all), as against 30.7% of men, while men were rather more likely to support Ramaphosa. However, large numbers (26% and 29%) of both genders supported no candidate.

We then asked voters whether they thought race relations in KZN had changed in recent times. Some 67.8% of ANC voters said that they had got much better or a bit better – an astonishing figure given the year-long campaign against “white monopoly capitalism”. There is no doubt that this has alarmed whites and Indians particularly – less than 2% of either group thought that race relations had got much better. One can only conclude that ANC voters wanted to give an optimistic answer and were blithely disregarding the actual effect of recent ANC propaganda on racial minorities. Only 10.9% of ANC voters thought that race relations had got a bit or a lot worse. By contrast, no less than 31.8% of Opposition voters took such a view.

We then asked how ANC voters how they would like to see ANC economic policy evolve in the future. The results showed that the great campaign for “radical economic transformation” has failed to convince even ANC voters. Among ANC voters, only 19.5% said they would like to see the ANC “push on with radical policies aimed at the complete redistribution of all wealth and income”. By contrast, 13.8% wanted the ANC to “carry on as now”, while 57.2% said they would like the ANC to “adopt more pro-business policies in the hope that business would invest more and create more jobs”. Even within President Zuma’s home province, ANC voters thus preferred a more conservative economic policy over a radical one by a 3:1 majority.

Finally, we broached questions involving state capture. Here, there was no doubt of the enormous damage done to the ANC by corruption and President Zuma’s involvement with the Guptas. Even among ANC voters, 32.5% thought that “South Africa is being run not by its government but by a criminal mafia which operates inside the government and also inside state-owned enterprises like Eskom and Transnet”. A further 17% thought there was “some truth” in that statement. Only 19.4% said there was “no truth at all” in that statement. Among Opposition voters the damage was even worse, with no less than 55.1% saying the statement was “completely true” and a mere 7.9% saying there was “no truth at all” in the statement.

We then pressed voters as to whether they thought there was “a corrupt relationship between President Zuma and the Guptas”. Even ANC loyalists in Zuma’s home province split down the middle on this crucial issue, with 41.8% saying that this was the case and 41.5% saying that it wasn’t. It is likely that at least some of the 16.6% of the ANC voters who refused to answer this question were cross-pressurised, not wishing to speak ill of Zuma but also unwilling to say that this statement was untrue.

Among Opposition voters the damage was even more patent, with 60.7% saying there was a corrupt relationship and only 13.9% saying there wasn’t.

We ended by asking voters what should happen in the event that “any corruption was proven between President Zuma and the Guptas”. Among ANC loyalists, 51% said that the Guptas should then go to jail, but Zuma should be pardoned. Only 9.8% thought that both Zuma and the Guptas should be pardoned, while 20.3% thought that both the Guptas and Zuma should go to jail. Some 18.8% refused to answer, doubtless again because of cross-pressurising effects.

Among Opposition voters the picture that emerged was even sharper. Only 31.6% thought that only the Guptas should be jailed, while no fewer than 39.5% wanted both Zuma and the Guptas to be jailed.

As regards the views of racial minorities, there were too few Coloureds in our KZN sample for us to make any comment on their perspectives. But KZN has one million Asians and nearly half a million whites. Both communities show signs of radical disillusionment, with Asians most affected. When asked whether they had voted ANC in either the last local or national election, 11% of Asians said they had, while 35% said they had voted for another party and 54% refused to say. The same pattern emerged when these groups were asked about the provincial government or the premier. Negligible numbers said that either had dealt with anything well, while around 60% of Asians (and rather fewer whites) either said they had dealt with badly with things or refused to answer at all. When it came to choosing ANC presidential candidates, 30% of Asians favoured Ramaphosa, while 10% supported Mathews Phosa and 50% supported no candidate. White responses were almost identical.

To summarise then, the Zuma forces, though doubtless at their strongest in KwaZulu-Natal, still commanded the allegiance of a little less than half of ANC voters. It also remained uncertain whether the current pro-Zuma provincial government would survive. This was nothing like as strong a position as President Zuma would doubtless have wished to see. At the ANC’s Mangaung national conference in December 2012, KZN delegates had voted for Zuma by 848 to nil, but the days of such impregnable solidity were clearly long gone. Already, no fewer than 32.9% of ANC voters would vote either for Ramaphosa or for Mkhize, as against the 44.3% that supported Dlamini-Zuma. This put the Zuma forces ahead by only a slim 4:3 majority in the run-up to the Nasrec gathering.

SURVEY RESULTS IN THE WESTERN CAPE

The Western Cape is politically different from South Africa’s other provinces in all sorts of ways. It is the only DA-controlled province (our respondents with partisan opinions were 58.8% Opposition and only 41.2% ANC supporters). It has by far the best known premier in Helen Zille (who was mentioned even in focus groups held in other provinces). It has the largest Coloured population of any province. The ANC vote here has three distinct components, with Africans comprising the overwhelming majority (79.5%), Coloureds constituting 18.2% of its support, and whites making up the other 2.3%. This African component is very largely Xhosa-speaking and is dominated by the residents of Khayelitsha. In previous opinion polls, this group has usually come across as the most alienated and radicalised section of the country’s African population: when US student Amy Biehl was murdered there in 1993, we found that 25% of Khayelitsha residents were sympathetic to the killers (which was more than twice the national average). ANC radicals like Winnie Mandela have always been more popular here than in any other part of the country.

By some margin, the Western Cape also has the lowest unemployment rate. In this province, moreover, ANC voters were evaluating a DA provincial premier and administration, which they often disliked for partisan reasons. By every benchmark, the Western Cape also has the most effective and least corrupt provincial administration of any in the country.

Also worth emphasising is the fact that the Western Cape has long experienced a large-scale and continuing inward migration to it from the Eastern Cape. As a result, Cape Town now has as many Africans as Coloureds. Ordinarily, one would expect this situation to produce a steadily rising ANC vote – whereas exactly the opposite has in fact occurred. This can only mean that the DA has successfully created a “DA political culture” in the province, which is either reversing or at least neutralising the political effects of that immigration. The DA’s successful administration of this province is thus working to a degree that Luthuli House must find provoking, even alarming.

Once we had omitted (as probable non-voters) those who answered “Don’t know” or refused to say how they had voted in the past, our sample was as follows: DA 51.1%, ANC 40.5%, EFF 4.6%, and Others 3.8%.

Our first surprise came when we asked voters how well they thought the province’s government dealt with corruption and found there was relatively little difference between ANC and Opposition voters on this issue. Some 36.2% of ANC voters thought the provincial government was doing very or quite well in this regard. Some 23.1% of ANC voters gave the same assessment as regards the provincial government’s success in alleviating poverty – which was hardly different from the 24% of Opposition voters who thought this. And there were actually more ANC than Opposition voters who thought the provincial government was dealing well with land reform and factionalism. This still meant, however, that only around a quarter of such voters approved of what it had achieved. On the factionalism issue, there was also a high (18.6%) “Don’t know” response, which might have reflected confusion as whether the factionalism in issue was within the ANC or the DA.

Much the same pattern was evident when it came to political violence, with 28.2% of ANC voters saying the provincial government was dealing well with the issue, compared to only 24.5% of Opposition voters – and with 15.5% of ANC voters responding “Don’t know”. These are better ratings than one normally finds ANC voters giving to ANC administrations in other provinces.

Even more striking was the good rating that the premier, Helen Zille, had from all voters. Among Opposition voters, 47% thought she was doing a good job and should stay on. A further 38% said she had done a reasonable job but they would still like a change, while only 10.7% said she had done a poor job and should go. Among ANC voters, 13.4% thought she’d done a good job and should stay on, whereas 40.5% said she’d done a reasonable job but they would still like a change. Only 36.7% of ANC voters thought she’d done a poor job and should go. These are remarkable figures, far better than any ANC premier achieved. Overall, the 53.9% of ANC voters who gave her a “good” or “reasonable” mark far outnumbered those who rated her “not good.”

If we look at Zille’s support in racial terms, we can clearly see how the DA political culture has spread across all groups. We thus find 56% of white voters saying she has done a good job and should stay in office, an opinion shared by 38% of Coloureds and 14.7% of Africans. A further 28.3% of whites said she had done a reasonable job but they would like to see a change, an opinion shared by 40.9% of Coloureds and 35.2% of Africans. Putting these two categories together, those who thought she had done a good or reasonable job included 84.3% of whites, 73.2% of Coloureds and 49.9% of Africans. This last figure is worth dwelling on: with half of all Africans having a good opinion of Zille, it is not surprising that the ANC has found it an uphill task contesting DA supremacy in this province. The DA’s real problem will be in finding someone who is as efficacious and popular as Zille when the time comes to replace her.

Something similar was visible as regards the question of how hard the provincial government has fought corruption. The number of ANC voters who said that DA provincial government was itself corrupt was astonishingly few – just 13.2% (less than the 18.2% of Opposition voters who thought that). No less than 43.9% of ANC voters said that the provincial government had fought corruption either “very hard” or at least “a bit” (much the same as the 43.5% of Opposition voters who said this). Yet in Cape Town there have been insistent rumours that property developers are bribing councillors to get planning permission. One can only conclude that the DA’s white electorate is more prone to believe such rumours than are the squatters of Khayelitsha.

In the run-up to the Nasrec conference, the Western Cape ANC made it clear that it strongly favoured Ramaphosa over Dlamini-Zuma. The preferences of ANC voters were much the same, with 41.9% of them preferring Ramaphosa as against 14.5% for Dlamini-Zuma, 11.7% for Mbete and 11.1% for Sisulu. Among Opposition (mainly DA) voters, Ramaphosa had 39% support, ahead of Sisulu (4.7%) and Dlamini-Zuma (2.7%). Some 37.9% said they had no preference, presumably because they supported no ANC candidate.

The question whether race relations has got better or worse in recent times is a politically loaded one in the Western Cape, where the ANC insists at every juncture that the DA is a white racist party and Cape Town is a racist city. Yet the results here are the opposite of what one might expect. No fewer than 51.4% of ANC voters said relations between the races had got a bit or much better. Only 3.2% said they had got much worse, while another 8.3% that they had got a bit worse. On the face of it, this is an overwhelming rejection of their own party’s propaganda.

Opposition voters, on the other hand, took a less sanguine view, with Coloureds feeling marginalised in the labour market by affirmative action policies and whites negatively affected by the propaganda against “white monopoly capitalism” and a general climate of hostility towards them. Among whites, thus, 42.6% said race relations had got much or a bit better, while 12.2% said they had got “much worse” and another 18% that they had got “a bit worse”. Given the poor state of the economy and the unfavourable political climate, these are nonetheless more favourable/optimistic figures for the province as a whole than might have been expected.

The results were also somewhat surprising when we asked voters in what economic direction they would like to see the ANC move – towards a more pro-business stance aimed at increasing business investment and jobs, or towards more radical policies aimed at the redistribution of all wealth and income. ANC voters here were far more in favour of more radical policies (37.8% favouring them) than they had been in KwaZulu-Natal. Given what we have said about the radical history of the local ANC, this was not unexpected. However, a surprising 27.7% of Opposition voters also favoured more radical policies – and this despite the fact that the DA accounts for over 86% of the Opposition vote and the EFF only 7.7%.

The ANC in the province is split down the middle on this issue, with 36.5% favouring more business policies, 16.2% saying “carry on as now”, and 9.6% responding “Don’t know”. One might have expected Opposition voters to be overwhelmingly in favour of more pro-business policies, but in fact less than half (47.3%) took such a view and a surprising 17.3% were “Don’t knows”.

When we asked whether voters believed there was a corrupt relationship between President Zuma and the Guptas, we found that there was almost no difference between Opposition voters and their ANC counterparts. Just over three-quarters of both groups (77.8% and 75.4% respectively) said they agreed that there was a corrupt relationship. Surprisingly, fewer ANC voters (8.9%) disagreed with this statement than did Opposition voters (12.2%). It is worth recalling that in KwaZulu-Natal, ANC voters were evenly split (42% to 42%) on this question. It is difficult to resist the conclusion that the very different results in the Western Cape reflect the fact that here Zuma is not protected by any favourite son/home province/Zuluspeaking bias.

Again, when we asked respondents whether they believed there was any truth in the idea that the South African state has been captured by a criminal mafia, there was surprisingly little difference between Opposition and ANC voters, with 48.6% and 41.9% respectively saying they agreed with the statement. By contrast, only 15.3% and 12.6% respectively said they thought there was little or no truth in the statement.

Finally, we asked what should happen if a corrupt relationship between Zuma and the Guptas were to be proven. Here there was a clear difference, with 25.4% of ANC voters but only 4.5% of Opposition voters saying that the Guptas should then be jailed whereas Zuma should be pardoned. Some 15% of both groups thought that both Zuma and the Guptas should then be pardoned, but 70.6% of Opposition voters and 47.5% of ANC voters preferred that both Zuma and the Guptas be jailed, if corruption were proven. Almost all voters in the province would thus be happy to punish a family of immigrant Indians, but a quarter of ANC voters would prefer to protect their party leader from such a fate.

In the run-up to the Nasrec conference, there was much talk about the votes of provincial delegates being manipulated. In the Western Cape, however, it seemed certain that the ANC’s provincial delegates would vote for Ramaphosa, as local ANC voters in the province also preferred.

SURVEY RESULTS IN LIMPOPO AND MPUMALANGA

Limpopo and Mpumalanga are both heavily ANC provinces, so much so that in both cases the number of Opposition voters in our sample was too small to be useful. Both provinces posed major questions in the race for ANC president. Mpumalanga, the fiefdom of David Mabuza, who has been premier for eight years, was long regarded as a lynchpin of the “premier league” supporting President Zuma. However, as the Nasrec conference came closer, Mabuza became ambiguous about which camp he supported, making him a potential king-maker. Mabuza also made it clear that he intended to become one of the ANC’s Top Six and perhaps even its deputy president. His position was made even more potent by a huge drive to increase ANC membership in his province, which gave Mpumalanga the second highest number of delegates to the ANC conference, behind only KwaZulu-Natal.

Other questions also arose regarding these provinces. Ramaphosa comes from a small Venda-speaking group in the north of Limpopo, so how far would he be able to convince the Northern Sotho (Pedi) majority of his province that he should be their candidate? Until recently the dominant figure in this province had also been Julius Malema. At the same time, Limpopo premier Stanley Mathabatha (who has held this post since 2013) is a reformer who has taken a strong line against municipal corruption. His open support for Ramaphosa resulted in an attempt to overthrow him by the pro-Zuma forces in the province, led by the ANC’s provincial secretary, Nocks Seabi. Indeed, the leaders of the pro-Zuma faction had met with Zuma in Nkandla to discuss how to get rid of Mathabatha and who his replacement should be. Mathabatha had managed to hang on, but there were doubts as to how far his public support in the province now extended.

Mpumalanga, on the other hand, is notoriously corrupt, a reputation which naturally clings to Mabuza as well. It was generally assumed that he had complete control over Mpumalanga’s ANC delegates to the Nasrec conference, but again there were questions as to the extent of his popular support.

We began by asking “How well has your provincial government dealt with:

In neither province do ANC voters have a very positive view of how far their provincial government is dealing with these issues, but Limpopo consistently scores more positive views and shows fewer negative ones. Given the heat of factional struggle within Mpumalanga, it is also worth noting that a record number of ANC voters (34.3%) said “Don’t know” or refused to answer our question about factionalism. This probably reflects a popular wish not to be involved in, and perhaps a fear of, such struggles. It is also worth noting that in Mpumalanga a record number of voters (38.6%) refused to answer our question about which party they had voted for. This was far higher than in any other province and recalls to mind the similarly high nonresponse rates we used to see in the parts of KwaZulu-Natal, where ANC vs. IFP fighting was particularly intense. It is certainly not a healthy sign.

We then asked voters how they viewed their respective premiers – David Mabuza (Mpumalanga) and Stanley Mathabatha (Limpopo): 

Again, these figures are a poor reflection on Mabuza. Only one in six ANC voters wanted him to continue as premier, while well over half those voters said they would like to see him step down. Moreover, the number of “Don’t knows” and refusals to answer is suspiciously high given that he has been so long in office. In Limpopo, by contrast, fewer ANC voters were unwilling or unable to hazard an opinion on whether Mathabatha should stay or go.

However, when we asked how hard ANC voters felt their provincial government had fought against corruption, Mpumalanga came out better than Limpopo. To be sure, only 2.1% of such voters said that Mabuza’s government had struggled “very hard” against corruption, against 10.6% in Limpopo, but more of Mabuza’s voters said his government had fought “a bit but not hard enough” against corruption (22.7% versus 12.9%). Moreover, 30.2% of ANC voters in Limpopo said that their provincial government “is itself corrupt”, as against only 20.6% in Mpumalanga. To be sure, Mathabatha inherited a government which was legendary for its tenderpreneurship practices, but whatever his good intentions he has clearly failed to dispel that reputation.

We next asked voters about their preferences for ANC president. Limpopo was a one horse race: Ramaphosa was the first choice of 77% of ANC voters, with Dlamini-Zuma coming second but a long way behind, with 3.3%. When we asked these voters for their second preferences, Ramaphosa, came in at 47.7% and was almost the only candidate mentioned. Indeed, Ramaphosa was clearly well on the way to making his home province a political base of some significance. Having such a base is important, as it seems to be a crucial determinant of political success in South Africa: Zuma (and Buthelezi before him) depended almost entirely on their base in KwaZulu-Natal, Mabuza on Mpumalanga, Magashule on the Free State, and Helen Zille on the Western Cape. President Thabo Mbeki’s lack of such a base was a fatal weakness.

Mpumalanga presented a far more dispersed and complex picture, though with Ramaphosa again well in the lead:

It was striking that Mathews Phosa fared so badly in Mpumalanga, where he was once the premier. In 1997 he had an approval rating of 62% in the province and was the only premier who also enjoyed widespread popularity in other provinces. In the run-up to Nasrec, however, he had been locked in battle with David Mabuza for some time and was clearly no longer a political force. It was also evident that ANC voters were by no means united behind Mabuza, but were instead very divided in their views. Contrary to common perception, Ramaphosa in fact had a large advantage here, though the Zuma faction had the support of just over one voter in seven. In Limpopo, by contrast, Mathabatha emerged as well justified in his pro Ramaphosa stance, for he was clearly reflecting the views of a large majority of his voters.

Next we asked voters whether relations between the races had got better or worse in their province in recent times:

In these two provinces, where the population is overwhelmingly African and ANC, there is a very strong optimism about race relations. This is particularly so in Limpopo. It is interesting to speculate whether this is connected to the province’s strongly pro-Ramaphosa views, for Ramaphosa is clearly favoured by most whites, Coloureds, and Indians. Given that our focus groups showed that most ANC voters thought the ruling party capable of self-correction once Zuma left – and since they usually wanted to get back to the “good old days” of the Mandela period – it is also possible that support for Ramaphosa is associated with the “rainbow nation” emotions of that era.

We then asked whether ANC voters would like to see the organisation move in a more radical or a more pro-business direction. Here, a sharp difference appeared between the voters of these two provinces:

As may be seen, ANC voters in Mpumalanga are evenly divided on which way to go, while in Limpopo an overwhelming majority would like to see the party move in a more pro-business direction. In neither case is there much support for current ANC policies. There is also a significant proportion of ANC voters in both provinces who find it difficult to give a clear answer to this question.

When we asked whether voters believed that the country had suffered state capture by a criminal mafia, popular feeling was again strongly one-sided:

The answers were similarly one-sided when we asked voters whether they thought there was a corrupt relationship between President Zuma and the Guptas. Some 61.2% of ANC voters in Mpumalanga and 67% in Limpopo thought that there was indeed a corrupt relationship, while only 12.5% and 8.2% respectively disagreed. The rest were “Don’t knows” or declined to answer.

Finally, we asked voters what should happen if such a corrupt relationship were to be proven.

These are extremely serious figures, which should give any ANC member or leader cause for reflection. Mpumalanga and Limpopo are both overwhelmingly ANC provinces – and yet, in both of them, two-thirds of ANC voters feel that there is indeed a corrupt relationship between the President and the Guptas – and that both of them should go to jail. This is an unprecedented situation: Mandela left office amidst universal approbation; Mbeki was voted out but no one thought he should be jailed; and Motlanthe was generally applauded after his short term. There has never been an ANC majority calling even for P W Botha or F W de Klerk to be jailed.

It should also be remembered that supporters of President Zuma are wont to point out that, while the ANC may have lost the major urban centres on Zuma’s watch, his support remains solid among rural voters and in the former Bantustans. The figures above show that this is not true: Mpumalanga and Limpopo are overwhelmingly rural provinces, while the KaNgwane “homeland” was in Mpumalanga and Venda, KwaNdebele, Lebowa, and Gazankulu were all in what is now Limpopo. The two provinces thus account for five of the ten former “homelands”, but two-thirds of the ANC voters within them nevertheless think that Zuma should be jailed if a corrupt relationship between him and the Guptas should be proved.

SURVEY RESULTS IN THE EASTERN CAPE AND THE NORTHERN CAPE

The Eastern Cape and the Northern Cape are contrasting provinces, with the former having one of the biggest delegations to the Nasrec conference and the latter one of the smallest. But both of them have seen bitter factional struggles of a very similar kind. In both, their premiers – who have been appointed by Zuma and are vulnerable to dismissal by him – have been easily recruited to the Zuma cause. However, both have also been opposed by pro-Ramaphosa majorities and their factional rivals.

In the Eastern Cape, the premier, Phumulo Masualle, who has been the ANC’s provincial chairman for the last eight years and assumed the premiership in 2014, was determined to win a further four-year term as chairman. Initially, Masualle was extremely critical of Zuma, but the Zuma forces – in the persons of Andile Lungisa (a key member of the ANC provincial executive) and Mzwanele (Jimmy) Manyi, who had just bought the Guptas’ New Age newspaper on very favourable terms – intervened strongly to turn Masualle around, no doubt telling him that he would be sacked as premier if he failed to back Dlamini-Zuma. As a result, Masualle quickly back-tracked and became a firm supporter of Dlamini-Zuma. This brought him into open conflict with the ANC provincial secretary, Oscar Mabuyane, who backed Ramaphosa, and who then stood for the ANC chairman’s post against Masualle.

Masualle, worried that he might lose at the provincial party conference, kept postponing this event. However, after three postponements, Mabuyane successfully insisted that it must be held, as the provincial executive had completed its constitutional term and its continuation in office was clearly illegal. It was duly held at the East London Convention Centre on 30th September. The Masualle faction forced its way into the gathering, demanding that Masualle be re-elected as chairman. Mabuyane and his supporters stayed grimly seated with folded arms through four hours of floor demonstrations in favour of Masualle. These ended in a physical brawl, with chairs thrown, at least eight people seriously injured, and the foyer of the Convention Centre covered in blood. As Mabuyane put it, the old leadership “just turned out to be hooligans”. At the end of this episode, the conference elected Mabuyane as the new chairman by 935 votes to 7. However, Mabuyane was then quick to say that the new provincial executive elected with him had no wish to prevent Masualle continuing as premier until 2019.

Masualle nevertheless attempted to get the courts to annul the conference. When this failed, he appealed to Luthuli House. But Mabuyane (a former SRC president at Fort Hare) pointed out that the conference had been overseen by members of the ANC’s national executive committee (NEC) and that its proceedings had been fully in line with ANC policy. He claimed that Masualle had attempted to collapse the conference and that, had he succeeded, this would have been followed by the collapse of the ANC December conference itself.

This left Masualle hanging on as premier – thanks to presidential patronage – while still heading a small pro-Dlamini-Zuma faction and still hoping to influence the provincial delegation, despite a now clear Ramaphosa majority. However, his position was further weakened on the eve of the Nasrec conference when the courts dismissed his challenge and confirmed the legality of Mabuyane’s election.

As regards the Northern Cape, Sylvia Lucas has been premier since 2013. She first hit the headlines as “the Fast Food Premier” as she had spent R53 000 on fast food in her first ten weeks in office – no less than R760 a day. Lucas is clearly a person who seeks accommodation with power. She joined the ANC in 1992, straight from her previous job as a secretary for the National Party. As premier, she retained the notorious John Block as her MEC for finance. She also quickly lined up with the ANC Women’s League in endorsing Dlamini-Zuma, thus earning President Zuma’s favour.

However, as the provincial party conference in May 2017 came closer, it was evident that the proRamaphosa forces, led by Zamani Saul and Desh Nxanga, would win. Lucas withdrew from the contest for the party chairmanship, rather than risk a public humiliation. Saul became the provincial chairman and Nxanga the provincial secretary. This was, naturally, a most displeasing result for the Zuma faction so, on the day before the conference was to be held, Lucas carried out a reshuffle of her provincial cabinet, purging it of pro-Ramaphosa elements.

Saul and Nxanga naturally cried foul: the provincial ANC conference had just endorsed Ramaphosa and made this decision binding on all ANC members and party structures. They pointed out that provincial cabinet changes could be made only after consultation with the provincial executive committee (the PEC) – and that Lucas had avoided all such consultation. The PEC even debated whether they should vote for Lucas’s dismissal as premier. This was a delicate question, given that the appointment and dismissal of premiers is a presidential prerogative. In the end the compromise reached was that Lucas would stay on as premier, but would have to reverse her reshuffle, as she then did.

This left Lucas in the same position as Masualle in the Eastern Cape: as a pro-Zuma premier hanging on in the face of a pro-Ramaphosa majority. Naturally, Lucas appealed to Luthuli House to nullify the provincial party conference that had brought her opponents to power. When that failed, she threatened legal action to the same end. Neither stratagem worked.

We began by asking voters how well their provincial government had dealt with the problem of corruption. As may be seen, in both provinces there was a deafening majority saying that the issue had not been satisfactorily dealt with:

A similar picture emerged when voters were asked how their provincial governments had dealt with the issue of poverty: 

An only slightly better picture emerged when voters were asked how well their provincial governments had dealt with land reform: 

It was noticeable that Eastern Cape ANC voters showed a somewhat more favourable view of their provincial government on these questions than their Northern Cape counterparts. This divergence was further amplified when they were asked how well their provincial governments had dealt with factionalism, with 45.4% of Eastern Cape voters saying their government had dealt very or quite well with that issue, compared to only 21.3% of their Northern Cape counterparts. The same picture emerged when voters were asked how well their provincial governments had dealt with political violence. Among Eastern Cape ANC voters, 52.1% said their government had dealt very or quite well with that issue, almost double the figure of 26.6% for their Northern Cape counterparts.

These outcomes were matched by a far better rating for Masualle as premier compared to Lucas:

Similarly, when we asked voters whether their provincial governments had made any effort to fight corruption, 15.2% of Eastern Cape ANC voters said their government had fought corruption very hard, compared to only 3.9% of their Northern Cape counterparts. Overall, however, ANC voters in both provinces – in common with ANC voters everywhere – were disappointed by the anti-corruption efforts that have been made. This issue also cuts right across party allegiances, with large majorities on every side wanting to see a far more vigorous pursuit of corruption.

When we asked ANC voters in the Eastern Cape who they would like to see as ANC president, only three candidates had significant support:

In the Northern Cape only two candidates had significant support: 

Dlamini-Zuma’s better showing in the Eastern Cape may have been a reflection of Masualle’s popularity being greater than that of Lucas. Even so, the choice was emphatic in both cases – 4 to 1 in the Northern Cape, and better than 2 to 1 in the Eastern Cape in favour of Ramaphosa.

When we asked about the state of race relations in recent times in these two provinces, there was no doubt that ANC voters in the Eastern Cape took a far more sanguine view:

Much of the divergence may stem from significant racial differences between the predominantly Coloured Northern Cape and the predominantly African Eastern Cape.

When we asked for views on the economic policy direction the ANC should take in the future, the verdict was deafening in both provinces: 

We next asked voters how they felt about a statement – “believed by some people” – that a criminal mafia was operating both inside the government and inside the state-owned enterprises (SOEs). As may be seen, more than 60% of ANC voters in both provinces thought this statement was partially or completely true.

We then asked voters whether they thought there was, as has been alleged, a corrupt relationship between President Zuma and the Guptas. The verdict was again deafening: 

Finally, we asked voters what should happen in the event that a corrupt relationship between Zuma and the Guptas should be proven. Again, the answer was emphatic: 

As may be seen, the opinions of ANC voters in these two provinces are not good news either for Zuma himself or for Dlamini-Zuma. In both provinces, the Zuma forces managed to line up the premiers on their side, but this merely produced successful rebellions by the pro-Ramaphosa forces. These insurgents also accurately reflected the opinions of ordinary voters in these two provinces. ANC voters did not merely reject the President’s nominee for ANC president: they also comprehensively disagreed with his policy of “radical economic transformation”. In addition, many were convinced that Zuma had allowed the corrupt capture of the state and that he should be punished accordingly.

To be sure, the Northern Cape had too few delegates to the Nasrec conference for its voice to carry much weight. But the Eastern Cape is a key province with considerable political clout. The survey results confirm that voter opinion there, as in the Western Cape, Mpumalanga and Limpopo, had moved strongly against President Zuma. If delegates were to vote at Nasrec according to the views of ordinary ANC voters, then only KwaZulu-Natal was likely to endorse Dlamini-Zuma at the Nasrec gathering.

SURVEY RESULTS IN THE FREE STATE AND THE NORTH WEST

Both the Free State and the North West are central parts of the “Premier League”, which was once quite solidly behind President Zuma, but later became greatly affected by splits in KwaZulu-Natal and by David Mabuza’s balancing act in Mpumalanga. Both provinces are ruled by strongman party bosses: Ace Magashule in the Free State and Supra Mahumapelo in the North West.

In the Free State, Magashule has been a prominent ANC leader since the early 1990s, but Thabo Mbeki was clearly distrustful of him. Hence, despite Magashule’s substantial local backing, Mbeki was careful to prevent him from becoming the provincial premier. Thus, while Magashule had been the chairman of the Free State ANC since 1992, it was only with the accession to power of Jacob Zuma in 2009 that he became premier. This helps explain Magashule’s strong allegiance to Zuma.

Moreover, as Zuma became increasingly entangled with the Gupta family, so too did Magashule. Two of Magashule’s sons work for the Guptas, and the Magashule family has reportedly benefited from a continuous series of favours, seemingly granted in return for the premier allowing the Free State to become part of the Gupta commercial empire. The Magashules have frequently gone on holiday with the Guptas and have allegedly benefited from foreign trips paid for by the Guptas. In return, Magashule is said to have allowed the Guptas free use of official Free State-headed notepaper with which to commandeer contracts, tenders and other goods. In addition, a dense network of scams seems to link the Gupta and Magashule families, while the Guptas have intervened in diverse ways – including buying airtime for hundreds of Free State ANC officials – to consolidate Magashule’s support. It has also emerged from the leaked Gupta emails that a large amount of public revenue, which had been allocated to start a dairy in Vrede, was actually used to pay for a Gupta wedding in Sun City in 2013. Naturally, Magashule denies that state capture exists or has occurred.

David Mabuza’s move to a more ambivalent position saw Magashule apparently waver in his commitment to Zuma. He emphasised his own role in helping Ramaphosa to become the ANC’s deputy president at the Mangaung national conference in December 2012. He also spoke in favour of a “unity ticket”. However, there was never really any doubt before the Nasrec conference that he would throw his support behind Dlamini-Zuma.

In the Free State, Magashule had long been dubbed “the life-time chairman”, but in time he found himself facing a serious challenge from his deputy-chairman, Thabo Manyoni. The two men fell out when Manyoni came out strongly against corruption and state capture and backed Ramaphosa for the ANC presidency. Magashule then deposed Manyoni as the mayor of the province’s biggest city, Mangaung, replacing him with a loyal Zuma-ite, Olly Mlamleni, who was the head of the local chapter of the ANC Women’s League. Manyoni responded by openly attacking Magashule and claiming that under him the Free State had become one of the most corrupt provinces. Manyoni also rather pointedly resigned as an ANC MP just before a no-confidence motion against President Zuma was put to the vote in Parliament in August 2017. Manyoni clearly seemed unwilling to vote against the motion, whereas Magashule demanded that those ANC MPs who had voted in favour of it must be disciplined.

The emergence of Manyoni as a serious challenger introduced a major new chapter in Free State politics: Manyoni commands considerable popular support – most estimates suggested a 50:50 split between him and Magashule. Unsurprisingly, the Free State’s ANC provincial conference was postponed, accompanied by the usual problems of verifying branch membership and status as the contest heated up.

It was also evident that Magashule relished the idea of joining the ANC Top Six, perhaps by replacing Gwede Mantashe as secretary-general. However, his faction was concerned that, if Magashule then had to quit as provincial premier, this could leave a clear field for Manyoni to romp home. Magashule thus temporized on this question, but clearly hoped that a Dlamini-Zuma victory at Nasrec would not only secure his ascent to the Top Six but also bring friendly forces into Luthuli House so as to ensure that Manyoni would be kept out.

In the North West, by contrast, provincial premier Supra Mahumapelo has remained a very forceful man who utterly dominates his province: his critics call him a dictator and claim that he is extremely corrupt. He managed to force out Thandi Modise as premier in 2014, largely due to his outspoken support for President Zuma. He remains strongly committed to the Zuma camp and is happy to describe himself as a friend of the Guptas. He completely denies that state capture exists and describes President Zuma as “a highly principled man”.

Mahumapelo’s muscular style has provoked outspoken opposition, with a major protest march mounted against him in January 2016 by ANC dissidents. This was led by the grand old man of North West politics, Rocky Malebane-Metsing, who had led the coup which deposed Lucas Mangope as president of Bophuthatswana in 1988. The protestors demanded that the ANC’s National Executive Committee (NEC), which was then sitting in Rustenburg, should remove Mahumapelo and dissolve his PEC. They said Mahumapelo had simply imposed candidates of his own choice, disregarding the elective principle. Despite the fact that the marchers had applied for and obtained official permission for their march, it was violently disrupted by young men in ANC T-shirts who hurled rocks and stones at them. These youths, it was claimed, were recruited by Mahumapelo and paid by him to disrupt the march. The youths also attacked journalists covering the march, apparently in order to protect their own anonymity. As a result the protestors were unable to deliver their memo to the NEC.

Mahumapelo describes his style of administration as “Supranomics”, which his critics claim is just a cover for looting public money – for he has quickly become the wealthiest of all the provincial premiers. Under him, the budget of the premier’s office increased by 87% in two years. In the run up to the Nasrec conference, he made it plain that he was not interested in a job in the Top Six, preferring to remain in place as premier. He claims to have a special vision for the North West (involving agriculture, culture, and tourism) and wants to rename the province after Moses Kotane. He interpreted in Setswana for Dlamini-Zuma when she visited the province, and declared that not only she – but all other candidates for office in the ANC – should be elected without a contest.

The North West is, of course, still marked by the events at Marikana in August 2012. In the local elections, the EFF emerged as the lead party there, while it scored well in the North West generally. ANC leaders, including President Zuma, have stayed away from Marikana, fearing hostile demonstrations. On her visit to the province, Dlamini Zuma had planned to lay a wreath for those killed at Marikana, but she was prevented from doing so by the now dominant mining union, the Association of Mineworkers and Construction Union (Amcu).

Mahumapelo’s critics claim that the North West ANC is in utter crisis – and this does not seem too strong a word. By August 2017 all the party’s provincial structures were heading back to court after they had been interdicted by ANC members who claimed that Mahumapelo was systematically setting up pliable and parallel party structures to replace the branches which opposed him. Already, they said, there were at least 40 such parallel branches. In addition, Mahumapelo’s PEC had dissolved three regional structures in the last two years. It was also allowing a fourth region to continue, even though its executive’s mandate had expired a year previously.

The August 2016 local elections gave rise to contestation within the provincial list committee, which was responsible for identifying the candidates to stand in the local poll. This committee was chaired by Sello Lehari, provincial MEC for Education and a key ally of Mahumapelo. Lehari reportedly used his powers to remove any dissident candidates, and then impose unpopular but loyalist candidates for both mayoral and councillor positions. Often, he allegedly went outside ANC structures to find people loyal to Mahumapelo. In order to guarantee his control, Mahumapelo reportedly insists that all ANC officials and representatives must sign an undated letter of resignation before they take office, so that if they cease to be loyal they can instantly be dismissed. The dissidence has been fiercest in Bojanala (where Rustenburg, the biggest town, is located), where the entire regional executive resigned in protest against Mahumapelo.

In October 2017 an ANC councillor in Mahikeng, Gaasite Legalatladi – who had been nominated for mayor in 2016 – was sentenced to five years in jail for plotting to kill Mahumapelo. Apparently annoyed by her failure to become mayor, she had visited a sangoma who told her that she was under Mahumapelo’s curse and that the curse must be reversed. Legalatladi then apparently asked for the sangoma’s help in assassinating both Mahumapelo and a former Mahikeng mayor, Gosianne Seatlholo. The sangoma reported her request to the police. Mahumapelo, for his part, knows that his dictatorial style often provokes bitter opposition and has publicly declared that he knows of nine different people plotting to kill him.

November 2017 witnessed another significant development, when Mahumapelo sought a court interdict against the “vicious smear campaign” allegedly being conducted against him by the leaders of the North West Business Forum, the province’s principal business organisation. The Forum is incensed by the murder of its late chairman, Wandile Bozwana, who was killed after he threatened to expose major public corruption. (Bozwana was shot nine times when a BMW M3 pulled level with his car and carried out a classic gangland execution.) Mahumapelo angrily rejects the implication that he ordered the murder, but the Forum’s leaders say they will continue to campaign against Mahumapelo’s corruption. They complain of his “unexplained wealth” and use adjectives such as “evil”, “greedy” and “corrupt” in describing him. The Forum welcomed Mahumapelo’s interdict application and said they would be happy to see him in court.

Despite these many challenges, in the run-up to the Nasrec conference, Mahumapelo continued to insist that his authority would prevail and that the North West ANC would support his chosen candidate, Dlamini-Zuma.

As the conference approached, the ANC in Luthuli House faced the largest number of branch disputes ever. This was clearly the result of ongoing and large-scale attempt to manipulate branches and impose candidates and delegates upon them. Luthuli House was repeatedly called upon to referee these bitter regional disputes, but was also seen as biased towards the Zuma forces. This apparent partisanship was greatly resented, particularly in the North West, where dissident ANC members felt that party headquarters was deliberately turning a blind eye to gross abuses. What was also apparent was that, whereas Mbeki had tried hard to prevent strong regional bosses from emerging, under Zuma such bossism had flourished.

Against this background, we began by asking ANC voters (the overwhelming majority in both these two provinces) how well they thought their provincial governments had dealt with corruption. Given the reputation for corruption that both provinces have, it was perhaps not surprising to find large majorities in both cases saying that their governments had dealt poorly with this issue.

Much the same was true when we asked ANC voters how well their provincial governments had dealt with poverty:

Again, much the same picture emerged when we asked ANC voters how well their provincial governments had dealt with land reform:

Much the same pattern emerged once more when ANC voters were asked how well their provincial governments had dealt with factionalism and political violence: 

***

When we then asked ANC voters to assess what sort of job their provincial premiers had done, the results were striking. Both men got a considerably better rating than many might have expected, but in the Free State opinion was equally divided as to whether Magashule should stay in office or quit, whereas in the North West two-thirds of ANC voters said they would like to see a change: 

When we asked ANC voters to judge how well their provincial governments had fought corruption, the results were mixed. In the Free State, opinions were dispersed right across the spectrum, with over half saying their government had made very little or no effort to fight corruption or was in any case corrupt itself. In the North West, opinion was decidedly more negative:

We then asked ANC voters to say who they would like to see as the next ANC president. In both provinces, the only significant scores were registered by Ramaphosa and Dlamini-Zuma.

These results were remarkable in several different ways. First, it was clear that both premiers intended to cast a united bloc vote for Dlamini-Zuma, yet in both provinces Ramaphosa was far ahead as the popular choice.

Secondly, in the North West no less than 41.9% of ANC voters refused to make a choice of candidates. This was an extremely worrying sign. Earlier, we reported how a general non-response rate of 38.6% in Mpumalanga resembled the figures we were used to seeing in KwaZulu-Natal at the height of the IFP vs ANC fighting, when voters were afraid to betray any preference. But the North West non-response rate here is even higher, suggesting that on this key question ANC voters were in a situation of extreme duress. This was presumably because many preferred Ramaphosa (or other candidates), but were under great pressure from Mahumapelo to support Dlamini-Zuma.

On some other questions, the non-response rate was also worryingly high in the Free State, particularly when we came to questions about corrupt networks, the Guptas, and so on. Voters were probably aware of the controversial links between Magashule and the Guptas, but preferred to avoid any open comment on such matters.

We then asked voters how they felt relations between the various races had fared in recent times:

In one sense, these results are unsurprising. Both provinces have overwhelmingly African populations and, while the racial minorities tend to believe that race relations have worsened, Africans have the opposite view.

We then asked respondents whether they would like to see the ANC move in a more radical direction towards the complete redistribution of all wealth and income, or whether they would like to see the party evolve in a more pro-business direction. In the North West an overwhelming majority took a pro-business view. In the Free State, opinion was far more scattered, but those preferring a more radical direction were the smallest group:

We then asked respondents what they thought of the idea that a criminal mafia was really running the government and the state-owned enterprises. The non-response rate was over 20% in both provinces on this question: almost always a sign that voters feel uncomfortable or even fearful about answering. The numbers giving credence to such a statement were thus far lower in these two provinces than in other parts of the country:

We followed up by asking whether respondents agreed or disagreed with the notion that there was a corrupt relationship between President Zuma and the Guptas. Majorities agreed, but again nonresponse rates were high:

Finally, we asked what respondents thought should happen in the event that a corrupt relationship between President Zuma and the Guptas was proven:

Again, though decisive, these results are less deafening than elsewhere.

One is left with a difficult picture of these two provinces. The large fact is that both are ruled by somewhat heavy-handed strong men, who were intent on delivering provincial bloc votes for Dlamini-Zuma, even though this was flatly against the preferences of ANC voters in both provinces. It is difficult to believe that this central conflict did not inform much of the atmosphere in which our other questions were answered.

For example, ANC voters were well aware that Dlamini-Zuma and the Zuma camp were campaigning for “radical economic transformation”, so voters would know that how they answered the question about the preferred economic policy direction of the ANC would have a direct bearing on their preferences for ANC president. This might explain the high (26.4%) non-response rate in the Free State on this question. Similarly, the bitter factional contest between Magashule and Manyoni almost certainly explains the high non-response rate in that province (at 29.2%) to our question about factionalism. In the North West, Mahumapelo’s determination to endorse Dlamini-Zuma almost certainly explains why non-response rates in the province rose to 41.9% when voters were questioned about their preferences for ANC president.

As we have seen, in both the Northern and the Eastern Cape, pro-Zuma premiers tried to push their provincial ANCs towards a vote for Dlamini-Zuma in the run-up to the Nasrec conference. In both these provinces, this pressure resulted in successful rebellions by pro-Ramaphosa factions which then came to threaten the premiers’ own positions. The Free State and the North West, on the other hand, provided a picture of provinces in which extremely dominant premiers had better prospects of defeating the proRamaphosa groundswell.

SURVEY RESULTS IN GAUTENG

Gauteng, the nation’s economic heart, is very much the home of the black middle class. Lacking either the Cape’s Coloured population or KwaZulu-Natal’s Indians, lower white collar and intermediate level employment is wholly in African hands. In addition, the concentration of (well-paid) African civil servants in and around Tshwane (Pretoria) is matched by the high number of black professionals and businessmen in Johannesburg. Necessarily, the ANC in Gauteng mirrors this fact. This is emphatically not a reliable pro-Zuma constituency: not only does the African electorate here include many of the “clever blacks” whom President Zuma has publicly disparaged, but in general the more cosmopolitan big city liberalism of Gauteng is at odds with President Zuma’s image as an uneducated, polygamous, and tribal Zulu.

It is thus no surprise that the Gauteng ANC has long been critical of Zuma. The dominant figure within the Gauteng ANC is Paul Mashatile, who has a long history of political activism in Johannesburg – he helped found the Alexandra Youth Congress back in 1983 (and was its first President) and played a prominent role in the United Democratic Front (UDF), also becoming general secretary of the Gauteng SACP. He became deputy-chair of the provincial ANC in 1996 and has been chairman since 2007. In 2008 and 2009, he was the provincial premier, stepping down only to become a cabinet minister. He held this position until he was dismissed by President Zuma by 2014 – probably as a result of Gauteng’s increasingly critical stance towards the President.

Mashatile’s dominance is such that all the other ANC leaders in the province are junior to him. The current premier, David Makhura, is Mashatile’s deputy-chair. Parks Tau, the mayor of Johannesburg from 2011 to 2016, was very much Mashatile’s junior in party terms. In Tshwane, Kgosientso Ramokgopa’s success in succeeding his aunt, Gwen Ramokgopa, as mayor in 2010 was generally considered a reward for his loyalty to Mashatile.

However, Mashatile’s position was severely affected by the August 2016 municipal elections. In the runup to the elections, the ANC’s national executive committee decided that Mashatile’s client, Kgosientso Ramokgopa, needed to replaced. It then imposed Thoko Didiza as its mayoral candidate. This caused enormous trouble from Ramokgopa’s clients and supporters, who fastened on to the fact that Didiza came from Durban and spoke neither Sotho nor Tswana. Five people died in the subsequent riots. In the event, the ANC lost control to the DA not only in Tshwane but also in Johannesburg and, briefly, in Mogale City (Krugersdorp). In addition, the ANC lost its overall majority in the province’s third metro, Ekurhuleni. Inevitably, Mashatile’s supporters blamed the unpopularity of Jacob Zuma for these results, while the Zuma faction blamed Mashatile.

The loss of Tshwane was highly symbolic – the ANC had campaigned frantically to prevent “regime change” in the nation’s capital. Still more importantly, the loss of Tshwane and Johannesburg has meant a huge loss of patronage for the ANC: a loss which becomes ever more severe as the two DA mayors, Solly Msimanga and Herman Mashaba, gradually weed out more and more ANC activists “deployed” to municipal posts. In addition, of course, municipal tenders and contracts in these two cities are now much less likely to be awarded to ANC-aligned interests.

The tension between President Zuma and the Gauteng ANC reached a high point in March 2016, when the Constitutional Court ruled that President Zuma had violated his oath of office over Nkandla. In response, the Gauteng ANC called publicly for Zuma’s resignation. Not long afterwards, when Zuma, Baleka Mbete and Jessie Duarte were all booed at May Day ceremonies, Mashatile commented that Zuma’s leadership had damaged the party. He warned that party members would have only themselves to blame if they did not recall Zuma, as the ANC might then lose the 2019 election. Mashatile also called openly for the ANC national executive committee to remove Zuma. At this juncture, it became obvious that Gauteng would support Ramaphosa in 2017 – and the ANC Youth League, which blamed Ramaphosa and Mashatile for the May Day booing, called for Ramaphosa to be booed on every subsequent public occasion.

After this, the Zuma faction openly campaigned for Mashatile and his executive to be removed. At an August 2017 meeting of the ANC national executive committee, the Zuma faction lined up Humphrey Mmemezi (a former Gauteng MEC) and Nomvula Mokonyane (a former Gauteng premier) to propose that Mashatile and his “ill-disciplined” executive should be disbanded. The motion failed to pass, but it was a major shot across the bows. The following month, when Makhosi Khoza resigned as an ANC MP, declaring her complete dissatisfaction with Zuma, Mashatile lamented that this was “a huge blow” to the ANC. By contrast, the Zuma faction in KwaZulu-Natal declared that this was a case of “Good riddance to bad rubbish”, and claimed that Khoza would now join “her masters” (the DA and Western imperialists). Perhaps even more dramatic was David Makhura’s announcement that the Gauteng ANC had considered the notion of “white monopoly capitalism” and decided that it simply didn’t exist. This slogan – in many ways, the centrepiece of the Zuma camp’s campaign – was just “a distraction”, he said.

The situation was also complicated by Mashatile’s ambitions. He had, after all, been continuously prominent in UDF/ANC politics for almost 35 years and he headed the region that included the nation’s political and economic capitals. So it was unsurprising that in August 2017 Mashatile declared himself a candidate for the ANC presidency. It was also quickly apparent that this was an over-reach, so he soon trimmed back to being a candidate for the Top Six.

Mashatile seems to have been greatly affected by David Mabuza’s sudden defection from the Premier League. This made Mabuza the “swing factor” that might decide the election and a potent candidate for the Top Six. Mashatile thus got together with Mabuza and began to argue, in unison with him, for a “unity candidate” – generally understood to mean Zweli Mkhize – rather than either Ramaphosa or Dlamini-Zuma. Mashatile even joined the so-called United League, which wanted an uncontested leadership election as the only way to avoid a split. This then put Mashatile together with Ace Magashule, Supra Mahumapelo and KZN’s Willies Mchunu, their watchword being that “ANC delegates must read from the same script”. Their approach, however, was badly outdated. The whole thrust of the 2017 presidential race within the ANC had been towards a more open and competitive contest – and the Premier League bosses were mistaken in thinking they could somehow shoe-horn delegates back into the patterns of the past.

In September 2017 Mashatile rowed back from his initial endorsement of Ramaphosa, declaring that this choice was “not cast in stone” and that “we’re ready to be persuaded”. This rather assumed that, if he was made a good enough offer, he could then deliver the Gauteng vote. But this was clearly not the case.

The fact remained that the Gauteng ANC was siding heavily with Ramaphosa. When the West Rand ANC branches were balloted, for example, they went 33 to 1 for Ramaphosa over Dlamini-Zuma. Branch officials did not appreciate Mashatile’s consorting with the likes of Mabuza, Magashule and Mahumapelo, nor his edging away from Ramaphosa. Open threats were made that Gauteng branches might not support his bid for a Top Six post. There was no resisting this and the Gauteng provincial executive committee then settled definitively on Ramaphosa as its candidate.

In probing voter views, we began by asking respondents how well they thought their provincial government had dealt with (a) corruption and (b) poverty:

As elsewhere, these ratings were predominantly negative though the score on “dealing with poverty” is rather better than in many provinces.

We followed up by asking how well voters thought their provincial government had dealt, respectively, with (a) land reform (b) factionalism and (c) political violence:

As may be seen, the ratings are very similar for all questions. As in every other province, opinion is negative on the whole. However, it is difficult to know how far this was determined by perceptions that the government in general is both corrupt and incompetent. Such feelings about the national government have probably strongly coloured many people’s views of their provincial government since, for many people, the two things are not strongly differentiated. As many see it, the government is simply “them”, “the elite”, “the authorities” or “politicians”. It is also worth remarking on the large number of voters who responded “ Don’t Know/No answer” to our question about political factionalism. This is a subject to which we will return.

We then asked voters to rate the performance of the provincial premier, David Makhura: 

Makhura is probably the most moderate ANC premier in the country, so it is interesting to note that this assessment is rather favourable to him, with over 64% of ANC voters rating his performance as “reasonable” or “good”. However, general disaffection with all government means that a majority would still like to see a change. Again, the non-response level was rather high.

We then asked respondents for their first and second choices among the various ANC presidential candidates. There was a greater dispersion of preferences here than in other provinces, perhaps as a result of better information about the less favoured candidates.

If popular preferences were to count, then Mashatile’s attempts to line up the Gauteng vote behind a “unity” candidate, such as Zweli Mkhize, had very poor prospects of success. In practice, as in other provinces, the presidential race had settled down in the popular mind as a two-horse affair. Mkhize’s problem was simply that he entered the presidential race far too late, well after opinion had crystallized.

We then asked how respondents felt race relations had fared in recent times:

We then asked ANC voters how they would like to see the ANC’s economic policies develop in the future – whether these should shift towards a more radical policy of complete redistribution, towards a more pro-business stance, or should rather remain the same:

As in other provinces, there is little support for more radical policies. However, the majority in favour of a more pro-business stance is muted by a high non-response rate.

We then told respondents that “some people” believed that a criminal mafia was running South Africa and controlling the government and the state-owned enterprises, and asked them how far they believed such a statement:

Again, the verdict is considerably less deafening than in other provinces. This is interesting, as Gauteng is the country’s economic headquarters and one would expect voters there to be more sensitised to such issues. However, as with the previous question, the figures are much affected by a high non-response rate.

Similarly, we then told respondents that some people believed there was a corrupt relationship between President Zuma and the Guptas and asked them whether they agreed with such a notion or not. Again, the non-response rate was high: 

Finally we asked voters what they thought should happen in the event that a corrupt relationship between President Zuma and the Guptas was proven:

Here too the verdict, though clear, was less pronounced than elsewhere, with a high non-response rate.

We were intrigued by the high non-response rate to several questions, because the Gauteng ANC is not plagued by either the intense factionalism or the political violence which typically produces this pattern of response. So we went back to our interviewers and asked them for their impressions on why so many respondents in Gauteng had declined to answer various questions.

We were told that many voters didn’t respond to the question about factionalism simply because this had ceased to be an issue in the province. When it came to the question about the premier, the high nonresponse rate came from the fact that David Makhura had a low profile and was not widely known. In general, people were far more conscious of their city councillors. In addition, Mashatile and the mayors of Johannesburg and Tshwane occupied most of the limelight, with the premier some way behind. As for the questions about corruption (those asking about a criminal mafia and the Guptas), our interviewers said the mood of many respondents was one of hopelessness and no longer caring. Typical comments were that, 23 years into democracy, little had changed while any changes evident were certainly not for the better. Many felt that the government did whatever it liked and that ordinary people were powerless. Voters felt punch-drunk.

These last comments are interesting because they come from South Africa’s chief metropolitan area. Experience in other parts of the world suggests that such areas are politically often a little ahead of other regions in a country. If this is so, such disillusionment may in time become far more widespread.

Footnote:

1 Tshivenda, Tsonga, Siswati, isiNdebele and other smaller groups.

This article first appeared in @LIBERTY: The policy bulletin of the IRR No 1/2018 / January 2018 I Issue 36

This issue is published with support from the Friedrich Naumann Foundation for Freedom.