Bob Mattes compares how the parties have fared since the last Opinion ’99 survey
The effectiveness of the political parties’ final four weeks of electioneering can be tested through a comparison of results from the April Opinion ’99 survey (the last opinion poll carried out before the election) with emerging election results.
Assuming a turnout of 85% to 90% of registered voters, the April survey showed the African National Congress with at least 68%, the Democratic Party 7%, the New National Party 6%, the Inkatha Freedom Party 4%, and the United Democratic Movement at 3%. The Pan Africanist Congress, Freedom Front, African Christian Democratic Party and the Federal Alliance all looked likely to get at least 1% each.
An additional 8% of undecided likely registered voters were also still available for parties to “top-up” these totals.
The successes of the ANC campaign lay in the sharp increases in voter support it had already achieved between September 1998 and April 1999. With some important exceptions, it appears that over the last four weeks it successfully protected its voting base against poaching by other parties, got them to the polls, but added few undecided voters.
In the Northern Cape, however, it seems to have finally capitalised on the past four years of a very popular administration by Manne Dipico to win 64% of the provincial vote, a significant improvement on any previous survey results.
In Gauteng, it appears that the late replacement of Matole Motshekga with Mbhazima Shilowa energised the ANC’s provincial efforts, winning it 66%of the provincial vote, up sharply from the 58% of registered voters who supported it in April.
And in the Western Cape, its 41% was eight points higher than the 33% of registered voters it polled in April. The NNP also drastically improved its standing in the province by taking nine percentage points from the large column of undecideds.
The most important gains of the last four weeks, however, appear to have accrued to the DP and the IFP.
While all the results have yet to be counted, it appears that KwaZulu- Natal saw significant swings of voting support compared to apparent party standings as revealed in the April surveys.
The so-called “lie factor” was immediately trotted out to explain the apparent discrepancy. This term, however, is misleading.
A range of international polling experience in volatile politicalE climates suggests that people rarely “lie” in the sense of telling pollsters they support party A while they actually support party B. Rather, voters who think they are in the minority, and fear possible exclusion, intolerance or intimidation, will “hide” their preferences from pollsters by saying they are undecided, often perpetuating a “spiral of silence” that can yield major surprises once all the voters are counted on election day.
Thus, the fact that the IFP has won around 8% of the vote nationally and 40% in KwaZulu, while the April survey poll suggested 4% and 17% respectively, can be understood as partially a function of those IFP supporters too cautious to reveal their preferences to pollsters, but “coming out of the closet” on election day.
However, the fact that opinion polls over the past four years have consistently found more confirmed ANC voters than confirmed IFP voters in KwaZulu-Natal strongly suggests that there are also a lot of truly undecided voters in rural areas. Ultimately, they vote for the IFP because it is seen as the “least bad” alternative to an ANC that has been thoroughly discredited (or vilified – depending on where you stand) by the IFP and its allied traditional leadership structures.
But a range of opinion surveys in early 1999 confirmed a departure from these patterns of the past four years, finding a potential sharp shift away from the IFP and toward the ANC, especially in rural areas
Thus, the IFP appears to have regrouped valiantly in the past four weeks, not only helping get “hidden” IFP supporters out of the closet, but clawing back some voters who may have been leaning toward the ANC.
Activists in the province give the IFP credit for running a solid campaign on the ground, focusing a late advertising campaign around the personality of Mangosuthu Buthelezi. They also suggest that the ANC campaign in KwaZulu-Natal may have failed by focusing on national issues and ignoring issues specific to the province, such as traditional leadership or the positioning of the provincial capital. The return of Lionel Mtshali is also seen as a strategic bonus because of his superior ability to communicate with rural voters.
While the DP had already enjoyed a vastly improved position since mid-1998 at the expense of the NNP, it saved the coup de grce for the last four to six weeks, using its “Fight Back” strategy and the personal pull of Tony Leon to stick the sword into the carcass of the NNP. It won a disproportionately large share of undecided voters to get 10% of the vote nationally (compared to 7% in the April survey).
While the DP made broad advances among white voters in the platteland, it appears to have totally routed the Nats in white suburbs, the old backbone of the once impressive Nat electoral machinery.
Its 19% gathered in Gauteng was substantially higher than the 15% of registered voters it polled in April. When the DP returns large winning margins in Centurion, ne Verwoerdburg, you know an electoral earthquake has occurred. All this was accompanied by strong gains in Indian areas, and surprising inroads into relatively better-off coloured communities in the Western Cape.
Compared to its poll standings of April, the NNP doesn’t appear to have convinced many undecided voters outside the Western Cape, where it protected its coloured voting base admirably.
Underfunded, and late out of the starting blocks, the UDM had since mid-1997 squandered away a great deal of the initial support that it received after its formation.
However, a furious run in the past two to three weeks appears to have recouped a great deal of that initial support in Eastern Cape, where it corralled a chunk of undecided voters and took voters back from the ANC to receive 13% of the provincial vote.
Bob Mattes is manager of the public opinion service of the Institute for Democracy in South Africa. The opinions expressed here are his and not necessarily those of Idasa or Opinion ’99, the election research consortium of which Idasa is a part