Article Hegemonism, Dominance or Weak Opposition? The Partisan Situation on the Eve of South Africa's Second Election Campaign Robert Mattes Helen Taylor Cherrel Africa As South Africa headed into a campaign for its second democratic election, the only real question in the eyes of most political commentators and analysts seemed to be the size of the victory that lay in store for the governing African National Congress. Would it manage to repeat its 1994 total of 62 per cent of the vote? Would it go even further and capture a two-thirds majority, enabling it to change the country's new constitution at will? But what reasoning underlay the foregone conclusion of an assured ANC victory? Was it because South Africa is becoming simply another repetition of the 'African pattern' whereby liberation parties that win the first election obtain increasingly larger margins of victory from voters who are either uncritical or uninformed? Or was it becoming another case study of a 'dominant party system' with voters deeply committed to political parties, no 'undecided middle,' and thus little or no prospects for electorally induced change in government (Giliomee 1998:133)? Or, was the impending ANC victory simply a function of elections in deeply divided societies, whereby the party that has managed to project itself as the sole legitimate representative of the largest ethnic or racial group can count on almost automatic support from voters in that bloc as expressions of communal solidarity, and thus elections come to resemble ethnic or racial censuses (Horowitz 1985 and 1991; Schlemmer 1994; Johnson 1996a). The results of the first survey of Opinion '99 (a consortium of EISA, Idasa, Markinor and the South African Broadcasting Corporation) conducted TRANSFORMATION 38 (1999) ISSN 0258-7696 Mattes, Taylor, Africa in September 1998, gave no reason to doubt the certainty of ANC victory1. They did, however, produce a range of evidence undermining all the arguments typically wielded to explain elections and voting in South Africa. Based on a wealth of findings, it is clear that the ANC was in a very strong and enviable strategic position. However, it did not occupy an unassailable position and had many areas of weakness that couldpotentially be exploited by skillful opposition parties. In terms of likely voting behaviour, the ANC was dominant. In terms of many of the factors that shape how people vote, however, the ANC was far from hegemonic among the electorate. As of September 1998, the ANC's measured voting support among the overall electorate was at an all time low since the 1994 election (51 per cent). And its levels of voter identification had also reached a post-liberation low (35 per cent). Not coincidentally, public opinion was also plumbing new depths in terms of optimism about the country and confidence in the new government. Voter optimism about the future of the country had sunk to its lowest point we had yet seen in the new democratic order. Evaluations of the economy, especially optimistic expectations, were hitting new lows. And evaluations of government performance were extremely negative on the issues that voters were currently ranking as the most important ones facing the country. Yet the ANC still enjoyed hefty leads over the opposition. Several factors accounted for this. Even though many people had become increasingly pessimistic aboutthe future of the country, not everyone blamed the government for these developments. There were also many dimensions of performance along which voters gave the government positive ratings. And when factored together, this resulted in a strong, overall approval of government performance. Finally, few people saw a real alternative offered by any opposition party. Very small proportions feltthese parties represented all South Africans, felt they were believable, felt they were competent to govern, trusted them to govern, or trusted them to handle a range of key issues. Thus, even as government support had increased over the previous year and a half, total support for all opposition parties had remained stable with large increases in the 'don'tknow' or 'undecided' columns. Only nine per cent identified with any opposition party. Taken together, these findings revealed a fairly sophisticated electorate, one that was relatively engaged and discerning, critical in some aspects and satisfied in others. South African voters were certainly not as blindly loyal to the parties that they supported in 1994 as widely depicted. Also, one could see in these attitudes significant potential for electoral change. The electoral strength of the Hegemonism, Dominance or Weak Opposition? governing ANC was not foreordained by the racial or ethnic composition of the electorate. Rather, the governing party was in a very strong strategic position due in large part to its own solid performance, but also due in an important way to the fact that opposition parties had failed to present most voters with a credible alternative. Trends in Political Party Support Voting patterns While it was in an enviable political situation on the brink of the country's second election campaign, the government was not politically unassailable. Indeed, over the previous year and a half the ANC had lost a significant amount of its 'confirmed' support. From late-1994 through mid-1997,61 percentto 64 per cent of potential voters regularly confirmed (through a secret ballot administered by Markinor interviewers) that they would votefortheANC'ifan election were held tomorrow'. However, by September 1998, that number stood at just 51 per cent of the whole electorate. At the same time, there was no corresponding increase in voting intention for the opposition parties. Even while there had been real declines in confirmed ANC support, intention to vote for any opposition party remained exactly where it was in late-1994 (28 per cent). Furthermore, only nine per cent of the electorate identified with any opposition party. What did occur was a significant upturn in the proportions of those voters that did not know who they would vote for, standing at 21 per cent of all potential voters in September 1998. Gra p h 1:Party Support 1994-1998 Source : O pinion 99, survey 1 (Septem ber 1998) n = 2200 70 4 —r ?4 . m 60 - 50 • ANC - — ^ ^ S 1 —m— O pposition 40 . Don't Know 30 - - ?n . 27—m-& B — - - 1 * — . - * « - — ~ * £ 20 . 21 16 14 12 12 12 10 - :o „• 11- 11 0 - Sept- May- Nov- May- Nov- May- Nov- Mar- Jul-98 Sep- Oct June 95 June 96 June 97 98 98 1994 1995 1996 1997 Mattes, Taylor, Africa Party identification If there is one thing that we have learned about voters in the past 30 years of election research is that most voters do not make up their minds anew in each election, nor do they make up their minds only when they enter the voting booth. Most voters tend to have preexisting predispositions about which party they 'identify with' or feel close to, predispositions that underlie actual voting choices(Campbelletal 1960;Key 1966;Fiorina 1979; Achen 1992;Popkin 1994). Voting scholars originally envisioned partyidentificationasalargelyunchanging attitude based almost solely on socialisation from parents and peers (Campbell et al 1960). However, we now know that 'partisan identification' is better characterised as a voter's 'standing decision' based on their evolving evaluations of political parties, leaders and performance (Key 1966; Fiorina 1979; Achen 1992). In the business world, most consumers tend to have standing choices of brands of automobiles, detergent or beer, based on whatever experience they have with those products, as well as on what information they can glean from advertising and trusted friends, family or colleagues. Much in the same way, a voter's standing decision is based on whatever information they can gather about political parties and candidates. At the age of 18, the best source of such information may indeed be one's parents or peers. However, that original decision is constantly updated by new information, information gleaned from the observations of economic trends visible in everyday life, from news media, or from respected friends and colleagues (Achen 1992; Popkin 1994). Voters whose experiences have led them to adopt such a 'standing decision' and identify with aparty are much more likely to vote, and very likely to vote for that party. However, just as there are consumers with no standing 'branr1 loyalty' and who are more open to competitive advertising messages or brand labeling on the store shelf, there are also many voters who, at any given point, do not identify with a political party. These' independent voters' are more open to the persuasion of an election campaign and may base their choice on party stances on the key issues of the day, as well as on how political parties and leaders campaign. At the same time, we know that political parties will have to work harder just to get 'independent' voters to actually come out and vote. In Idasa's first post-election survey conducted in September-October 1994, we found exceptionally high levels of voter identification with political parties. When asked whether 'regardless of how you voted on election day (in April) was there one particular party which you felt especially close to?' 88 per cent said they did feel close to one particular party. These high levels of voter Hegemonism, Dominance or Weak Opposition? identification with political parties seemed to be an important factor that distinguished South Africa from many other emerging democracies, such as Russia or elsewhere in Eastern Europe where parties are relatively new, often short-lived, and enjoy very low levels of voter identification (see White et al 1997) .Itis important to remember thatwith one exception(the African Christian Democratic Party), voters could point to a long history of prior experience with everyparty—or its previous organisational form) that ultimately wonparliamentary representation. Since then, however, this number dived rapidly, and accounted for the growing number of people who could not tell us how they would vote. By 1995, just 18 months after the 1994 election, only 5 8 per cent said they identified with a political party (it should be noted, however, that the question wording changed slightly: 'Do you usually think of yourself as close to any particular party/ies?') Indeed, a follow-up question found that over half of the 42 per cent who could be called 'independent' (or 22 per cent of the total sample), said that theyneitheridentifiedwithapartynor ieaned' toward any specific party (what we might call 'hardindependents').By 1997,56per cent said they didnot identify with any political party; 23 per cent also said they that they did not lean toward any party. As of September 1999, only 45 per cent identified with a party, and 36 per cent maintained that they did not lean in any direction. Graph 2: Identifiers vs Independents Source: Opinion 99, survey 1 (September 1998) n=2200 100 90 •M- «T°° 80 • _ Identifiers 70 «_ Independents 60 50 40 -I 30 20 10 0 Sept-Oct 1994 Sept-Oct 1995 June-July 1997 Sep-98 Mattes, Taylor, Africa While 1994 may be a demanding comparison, the 1994 results do elucidate the type of underlying attitudinal situation that produced the voting patterns ultimately seen on 27 April. The subsequent findings also demonstrate how much the environment of April 1994 differs from that of late-1998. Many analysts have called the 1994 election a 'racial' or 'ethnic' census (Schlemmer 1994;Giliomee 1994; Welsh 1994;Reynolds 1994;Sartori 1995). They have pointed to previous and ensuing opinion polls to suggest that voter behavior in South Africa is driven largely by the desire to display communal solidity with political parties representing your community, or out of a fear of communal pressures and intolerance (eg Johnson 1996b; Johnson and Schlemme, 1996; Goetz and Shaw 1996; and Johnson 1999). But developments in partisan identification indicate that South African voters are not nearly as blindly loyal to their political parties as the 'census' interpretation would suggest. In addition, a wide range of other data from 1994 post-election surveys seriously question many of the implied hypotheses of the racial census explanation of voting in South Africa (see Mattes 1996; Mattes and Gouws 1998; and also Seekings 1997). Graph 3: Party Identification 1994-1998 Source: Opinion 99, survey 1 (September 1998) n=2200 70 60 50 40 30 20 10 0 Sept-Oct Sept-Oct June-July Sep-98 1994 1995 1997 Hegemonism, Dominance or Weak Opposition? In terms of identification with any specific political party, voter identification with the ANC in September 1998 stood at slightly more than one-third (35 per cent), a sharp drop from the 58 per cent registered in the 1994 post-election survey. At the same time, identification with opposition parties had also dropped dramatically. As of September, only nine per cent of the electorate identified with any opposition. The former occupants of power, the National Party, saw their levels of identification crash from 15 per cent to three per cent in the space of four years. Factors Influencing Trends in Party Support Why had ANC support dropped? A useful indicator that summarises a broad range of electorally relevant voter sentiment is a simple question about whether the country is headed in the right or wrong direction. Since 1994, when 76 per cent of a national sample said the country was headed in the right direction, regular Markinor surveys tracked a gradual decline in optimism about the overall direction of the country, with especially sharp drops after 1996. The first Opinion '99 survey registered the first time that the optimistic and pessimistic trend lines have actually crossed, with 44 per cent saying the country was headed in the wrong direction, and 43 per cent saying right direction. While the drops had been especially sharp among white, coloured and Indian voters, there had also been important drops among black voters (with only 53 per cent saying 'right direction'). Graph 4: Overall Direction of the Country 1994-1998 % saying heading in the right direction Source: Opinion 99, survey 1 (September 1998) n=2200 80 70 re— 60 50 40 30 20 10 Jun-94 Nov-94 May- Nov-95 May- Nov-96 Jun-98 Sep- June June 98 1995 1996 Mattes, Taylor, Africa At the same time, Idasa surveys over the previous few years had tracked progressive decreases in several different indicatorsof economic optimism. The proportions of voters who were satisfied with the national economy had moved downward between 1997 and 1998. The numbers of voters who feel that the economy was improving had also declined considerably over the previous four years.Even while surveys since 1994 had revealed low levels of optimismabout present or recent economic developments, they had also shown that South Africans had remained fairly optimistic about the future. Now this too was starting to erode. By September 1998, only 34 per cent felt that the economy would improve in the next 12 months. This drop was especially notable among black voters where 41 per cent maintained optimistic expectations, down from the 53 per cent of mid-1997. Graph 5: Evaluations of the National Economy Source: Opinion 99, survey 1 (September 1998) n=2200 50 44 40 30-j . Past Year 25 20 . Present 10 Future 0 1994 1995 1997 1998 Perhaps most worrying from the government's perspective was that its worst ratings came on many of the very issues that voters were identifying as the most important ones facing the country. Over the previous four years, jobs, education and housing had consistently received top priority from voters as the country's 'most important problems.' During this period, concerns about crime had skyrocketed as a public concern from single digits in 1994 to now rank as the second most important problem facing the country. With the drop in the value of the Rand, sharp increases in Hegemonism, Dominance or Weak Opposition? interest rates, and price rises in mid-1998, broad concerns over 'the economy' had also recently emerged to become ranked as one of the five most important. Thus while 73 per cent of the electorate cited jobs and employment as the most important problem for government to address, an abysmally low 12 per cent thought the government was doing a good job in this area. Sixty-four per cent cited crime, yet only 18 per cent approved of government performance in this area. And only 18 per cent felt the government had done a good job on the fast- emerging problem of the overall economy. But these results also indicated that voters were not simply ranking an issue as an important problem only when they were dissatisfied with the way government was addressing it. In many instances they could simultaneously feel that something was a pressing national problem and also feel that the government was doing the best job it could under the circumstances. For example, housing and education were the third and fourth most frequently mentionedproblems, respectively, yet approximately one-half of voters said the government was doing a good job in these areas. While government did receive much better ratings on other problems regularly probedby Markinor, they were not ones that voters thought were important such as water provision, ending political violence or nation-building. Graph 6: Government performance on Public Priorities Source: Opinion 99, survsy 1 (September 1998) n=2200 • Most lmportant| Problem I % saying govtis handling problem well Mattes, Taylor, Africa Why was the ANC still so far in front? Thus, as the country geared up for its second election campaign, 56 per cent of all potential voters said they did not feel close to any political party. Four- in-ten insisted that they did not even 'lean' toward any specific party. And one-in-five said they did not know who they would vote for if an election were held tomorrow. With so many voters without a 'standing choice,' one might have expected opposition parties to be in position to reap an electoral bonanza a few months down the line. Yet opposition party strategists had few reasons to rejoice at these otherwise favourable conditions. There were a series of important reasons why the ANC still enjoyed a commanding lead in popular support and that government losses had not turned into opposition gains. Some of them had to do with the government, and some with the complexity of governing a country like South Africa. But many of them had to do with the very nature of the opposition itself. To fully understand South Africa's partisan environment in 1998, it was not enough to know whether voters thought things were getting better or worse. It was also important to know whom they blamed or rewarded for the overall direction of the country and economy. In fact, even while voters increasingly saw the country going in the wrong direction, they did not necessarily blame the government for these developments. We asked people to tell us who they felt was responsible for the overall direction of the country: was government completely responsible, or were other factors partially, mostly or wholly responsible? Overall, 59 per cent said the government was either completely (49 per cent) or mostly (12 per cent) responsible for the direction of the country. Looking only among the politically vital black vote, just 53 per cent felt the country was going in the right direction by September 1998. As might be expected, 55 per cent of these optimistic black voters gave all reward to the government; 66 per cent gave all, or most of the reward to the government. However, among the one-third (33 per cent) of pessimistic black voters who thought the country was going in the wrong direction, only 37 per cent laid all the blame at the government's door. Forty-three per cent said 'other factors' were 'mostly' or 'equally responsible' for the country's downward trends. Read in another way, 52 per cent of these voters felt that someone else beside the government had to share in at least some, or take all of the blame. 10 Hegemonism, Dominance or Weak Opposition? Graph 7a Country going in right direction (Who is rewarded) Among black respondents (Sept 1998) Source: Opinion 99, survey 1 (September 1996) n=2200 fin 50. 40 30. •• 20. ^^™ 14 H ^m WM 8 f, 7 m 0 m ,• , m Govt Govt mostly Govt + other Other Other Dontknow completely responsible factors factors factors responsible mostly completely responsible responsible Graph 7b: Country going in wrong direction (Who is blamed) Among black respondents (Sept 1998) Source: Opinion 99, survey 1 (September 199B) n=2200 Govt Govt mostly Govt + other Other Other Dontknow completely responsible factors factors factors responsible mostly completely responsible responsible n Mattes, Taylor, Africa Consistent with these trends, we also found that even while voters thought the country was caught in a downturn most seemed to be satisfied with the overallperformance of the government. Since 1995, approximately eight-in-ten had consistently said that President Nelson Mandela had been handling his j ob 'well' or 'very well'; Deputy President Thabo Mbeki also received increasingly positive ratings (67 per cent in September). This bode well for the ANC since comparative experience from other democracies suggested that the popularity of outgoing or incumbent executives had an important influence on theelectoral fortunes of their party. Popular affirmation of performance, however, was not restricted to heroic personalities in the executive office. Overall government performance also received strong, though lower, levels of public approval (58 per cent). Yet even if voters saw things going in the wrong direction, and even if they blamed the government for this, voters would still have needed to see a legitimate, credible alternative in order to shift their votes. The September Opinion '99 survey posed voters with a wide range of policy issues and asked people which, if any, parties they trusted to handle this issue. Even on those key issues where dissatisfaction with ANC performance was widespread, few voters saw any other party as a real alternative. Thus, while only 12 per cent approved of ANC performance on the issue of job creation, only 24 per cent said they could trust the National Party to handle employment issues, 14 per cent said the Democratic Party, and even smaller totals mentioned any other opposition party. And while only 18 per cent approved of the way government was handling crime, just one-quarter said they could trust the NP tofightcrime, 13 per cent the DP, and 11 per cent cited thelnkatha Freedom Party. From the standpoint of the opposition, the picture was depressingly the same on all other issues that we polled. Across the 13 performance areas included in our questions, at no point did any more than one-quarter of the electorate say they trusted any opposition party on any issue. But a viable political alternative comes not just from political parties taking different policy 'positions,' but from convincing voters that they can trust you to handle new issues as they come up, to govern generally well, and to govern in their interests. Given the low amount of factual information that most voters possess about policy positions, or party records, voters often are left with inferring information about candidates andparties (what SamuelPopkin (1994) calls 'low-information reasoning'). Most importantly, voters try to figure out whether a party is credible and likely to govern well. In lieu of other information, they often do this by simply looking at: 12 Hegemonism, Dominance or Weak Opposition? • demographic characteristics of candidates and party leadership • whether respected community leaders might support the party • what knowledgeable friends and neighbours might be saying • where and how a party and its leaders campaign • past track records of parties and candidates. Thus, without any other information, voters look to things like skin colour, language, accent, who is leaving or joining a party, what people like them are saying about a party, and whether a candidate appears to know and understand their culture and what their problems are. We examined voterperceptions of all the major political parties across a wide range of criteria. We asked people whether: • they thought each party represented all South Africans, or one specific group? • they could believe what a party says? • they like what each party had done, as a party, over the previous four years? • they could trust each party to do what is right if it ran the government (provincial or national)? • and finally, whether they felt each party could do a good job running the government if it were elected? The resulting picture across all these criteria is resoundingly the same. On virtually every criteria positive perceptions of the ANC ran far ahead of any other opposition party. More importantly, perceptions of almost all opposition parties were fairly to exceptionally negative. The responses to these questions clearly demonstrated that the problem of the opposition was not simply a consequence of people who did not like or did not know the policy positions of opposition parties. The roots went much deeper. The 1994 Idasa post-election survey revealed that the overwhelming majority of voters supported parties that they felt were inclusive. The only voters who seemed to feel that their party was exclusive to their own group were PAC, FF and IFP voters (Mattes 1996). But, by 1998, it was clear that opposition parties had failed to use four years of participation in either the Government of National Unity or in Parliament to improve their images. For example, perceptions of the NP as an inclusive political party, standing for all South Africans, fell from 39 per cent to 32 per cent and for the liberal, centrist Democratic Party dropped sharply from 39 per cent to 24 per cent. 13 Mattes, Taylor, Africa Graph 8: Percentage who think these parties represent all South Africans Source: Opinion 99, survey 1 (September 1998) n=2200 90 80 77 —71 70 • 1994 • 1998 60 50 40 • 39 39 • 32 1k 1fe 30 2426 22 • I 20 20 10 8 6 0 ANC DP FF IFP NP PAC UDM With regard to party credibility, 57 per cent felt that they couldbelieve what the ANC said 'most' or 'all ofthe time' (not an especially high figure foragoveming party), the nearest opposition party was the NP at 16 per cent. In fact, four in ten voters felt that the NP (42 per cent) and the Inkatha Freedom Party (42 per cent) could 'never' be believed. Graph 9: Party credibility Source: Opinion 99, survey 1 (September 1998) n=2200 Bn. 57 50 40- ± • Credible always/ most of the time I 30- 16 20- nid 10 0- 4 J I— —m ANC DP 6 FF 10 FP NP 7TW 10 PAC 6 UDM 14 Hegemonism, Dominance or Weak Opposition? The pattern of responses was largely the same to a question about whether people felt they could trust each party to 'do what was right' if it were elected to run the government. And all of these images translated into very different images of competence to govern. Where over two-thirds of all potential voters said they felt the ANC could do a 'good' or 'very good'job running the national government if it were elected, 28 per cent said so about the NP, and 21 per cent for the Democratic Party. The figures were lower for all other opposition parties. Graph 10: Trust in political parties Source: Opinion 99, survey 1 (September 1998) 60 -, 57 I 50 • Trust them always/ 40 - most of the time 30 - 18 20 - 12 9 m 10 7 6 10 0 ANC n: DP • TTITI FF IFP NP PAC UDM Graph 11: Competence to Govern Source: Opinion 99, survey 1 (September 1998) n=2200 Very good /good ANC DP FF FP NP PAC UDM 15 Mattes, Taylor, Africa A clear pattern of party images emerges from these responses. Overall, the ANC received widespread positive images in September 1998, and was widely known. In contrast, the NP and Inkatha Freedom Party were also widely known, but enjoyed very negative images. The Democratic Party, Pan Africanist Congress, and Freedom Front received, on balance, negative responses from those who were able to rate them. But significant proportions of the electorate were unable to offer an opinion about them across any indicator (usually more than four of every ten voters). Finally, among those who knew something about the recently formed United Democratic Movement, the party received, on balance, favourable responses. But most voters (70 per cent to 80 per cent) were unable to offer opinions about the UDM. Electoral competition and democratic consolidation What do these initial results from the Opinion '99 project tell us about voters, voting, and the consolidation of democracy in South Africa? Most importantly of all, the findings strongly point out the need for analysts and politicians alike to rethink many of their simplistic understandings of South African voters. They are not nearly as blindly loyal or as unresponsive to 'political reality' as is widely assumed. This is true as well of black voters who have traditionally been portrayed by political analysts as solidly, even blindly committed to 'their' political parties, and lacking the ability or willingness to distinguish between positive and negative areas of performance. Voter support for political parties is not as nearly as committed, or 'glued in' as most observers have thought. When deciding which parties to support, they look to real-world events (or at least how they perceive them) such as the economy, government performance, and how the country is doing generally. Finally, when evaluating the economy, government and political parties, voters do not fall into a simple 'for them/against them' mindset. They are quite capable of distinguishing between, say how they are doing personally versus how the country is doing, or between how things are today versus how they expect them to be in a year's time. They are also quite able to differentiate between a range of dimensions of political performance, for example, between evaluations of government performance on housing versus crime, and whether a party is competent to govern versus whether they can believe what it says. With regard to the future of elections and political parties, it is clear that the ANC does not occupy a permanently unassailable position. It has, and most certainly in the future will continue to have, many areas of weakness 16 Hegemonism, Dominance or Weak Opposition? and vulnerability in terms of voter evaluations, areas that could potentially be exploited by skillful opposition parties. It is far from hegemonic in terms of public opinion. In this, one can see significant potential for electoral change. Yet with the possible exception of the fledgling United Democratic Movement, fewpeople now see any real alternative offered by any opposition party - certainly far too few in the ranks of those who supported the ANC in 1994. The measured decline in ANC support over the previous year and a half was a partial sign of the extent to which democratic politics in South Africa was 'normalising' by September 1998. Governments must make decisions. And while they hope to satisfy a majority in order to remain in power, the very act of governing would have almost certainly dissatisfied many people, including many of its original supporters. As Lincoln said, 'you can't satisfy all of the people all of the time'. Thus, with its support in the 1994 election and in subsequent opinion polls already soaring at stratospheric levels, a drop could have been expected. The prospects of a consolidated democracy in South Africa hinge greatly on the ANC government's ability to meet material expectations and improve people's lives (see Mattes and Thiel 1998). Butthe prospects of consolidation also depend on the existence of a viable alternative capable of providing a political home to those voters who are dissatisfied with government performance. Only in this way will voters in a young democracy be able to blame a specific party for any grievances they harbour, rather than the democratic system itself. And only eventually under different governments, and by voting both for winners and losers will they be able to distinguish clearly between political parties, government, and larger political institutions. At least in the minds of voters, it seems it is the opposition, more than the government that is not delivering on its part of this bargain. The real long- term threat to stable, competitive multi-party democracy now seems to be the lack of a credible, legitimate opposition. If left unchecked, over time, dissatisfied voters with no place to turn may become apathetic, may not care about the survival of democracy, and may become increasingly more discontented. Whether opposition parties were able to use the campaign for the June 1999 elections to change the way they are perceived by voters may tell us a great deal about the future of competitive multi-party democracy in South Africa. 17 Mattes, Taylor, Africa Notes 1. EISA is the Electoral Institute of South Africa and IDASA is the Institute for Democratic Action in South Africa. Opinion '99 results may be found at http:// www.idasa.org.za References Achen, C. 1992. 'Social psychology, demographic variables and the linear regression: breaking the iron triangle in voter research', Political Behavior 14/3. Campbell, A, P Converse, D Stokes and W Miller. 1960. The American Vote. New York: John Wiley. Key, V O. 1966. The Responsible Electorate: rationality in presidential voting, 1936-60. Cambridge: Harvard University Press. Fiorina, M. 1979. Retrospective Voting In American Presidential Elections. New Haven: Yale University Press. Giliomee.H. 1994. 'The National Party's campaign for a Liberation election'. In: Andrew Reynolds (ed), Election '94 South Africa: the campaigns, results and future prospects. Cape Town: David Philip. 1998. 'South Africa's emerging dominant party regime', Journal of Democracy 9(4). Goetz, G and M Shaw. 1996. "The election on the Reef: choice and first-time voters in Gauteng (the PWV)'. In: RW Johnson and L Schlemmer (tAs), Launching Democracy: South Africa 'sflrstopen election. New Haven: Yale University Press. Horowitz, D. 1985. Ethnic Groups in Conflict. Berkeley: University of California Press. 1991. A Democratic South Africa ? Constitutional engineering in a divided society. Cape Town: Oxford University Press. Johnson, R W. 1996a. 'The 1994 election: outcome and analysis'. In: R W Johnson and L Schlemmer (eds), Launching Democracy: South Africa 'sflrstopen election. New Haven: Yale University Press. 1996b. 'How free? How fair?' In: R W Johnson and L Schlemmer (eds), Launching Democracy: South Africa 'sfirst open election. New Haven: Yale University Press. and L Schlemmer. 1996. 'National issues and national opinion'. In: RW Johnson and L Schlemmer (eds), Launching Democracy: South Africa'sfirst open election. New Haven: Yale University Press. 18 Hegemonism, Dominance or Weak Opposition? 1999. 'Voting and the Fear Factor", Focus 13. Mattes, R. 1996. The Election Book: judgement and choice in South Africa's 1994 election. Cape Town: Idasa. and A Gouws. 1998. 'Race, ethnicity and voting behaviour: lessons from South Africa'. In: Sisk and A Reynolds (eds), Elections and Conflict Management in Africa. Washington: United States Institute of Peace Press. andHThiel. 1998. 'Public opinion and consolidation in South Africa', Journal of Democracy 9( 1). Popkin, S. 1994. The Reasoning Voter: communications and persuasion in presidential campaign. Chicago: University of Chicago Press. Reynolds, A. 1994. 'The results'. In: A Reynolds (ed), Election '94 South Africa: the campaigns, results and future prospects. Cape Town: David Philip. Sartori, G. 1995. 'Constitutional court needs minorities to protect democracy', Business Day, August 2. Schlemmer, L. 1994. 'South Africa's first open election and the future of its new democracy'. In: H Giliomee and L Schlemmer (eds), The Bold Experiment: South Africa's new democracy. Halfway House: Southern Publishers. Seekings, J. 1997. 'From the ballot box to the bookshelf: studies of the 1994 South African general election', Journal of Contemporary African Studies 15(2). Welsh, D. 1994. The democratic party'. In: A Reynolds (ed), Election '94 South Africa: the campaigns, results and future prospects. Cape Town: David Philip. White, S, R Rose and I McAllister. 1997. How Russia Votes. Chatham, N J: Chatham House. 19