In April 2025, the Institute of Race Relations (IRR) released polling data that has set tongues wagging across South Africa’s political landscape. For the first time, the Democratic Alliance (DA) allegedly edges out the African National Congress (ANC) in national support. On the surface, it reads like a game-changing moment. But a deeper look raises serious questions, not just about the findings, but about the messenger.
The poll shows the DA at 30.3% and the ANC at 29.7%, a slim but symbolically potent lead. The timing is critical: the survey was conducted during a national uproar over a proposed VAT hike that saw the ANC bulldoze its budget through Parliament, seemingly without full GNU consensus. The DA’s vocal opposition to the tax increase and its polished messaging during this crisis clearly won it some points with voters, especially black voters, if the poll is to be believed. According to the IRR, DA support among black voters rose from 5.4% to 18% in under a year.
The IRR is not just any polling outfit. Its leadership and affiliates have long blurred the lines between think tank and political shop. Helen Zille, the Chairperson of the Federal Council of the DA, served as a senior policy fellow there. Gwen Ngwenya, the DA’s former Head of Policy, was its COO. Gareth van Onselen, the DA’s ex-communications head, ran its Politics and Governance desk. Even the party’s youngest MP, Mlondi Mdluli, previously managed campaigns for the IRR. These are not passive alumni; they are ideological insiders.
This brings us back to the question: can the IRR credibly claim neutrality while its benches remain warm with DA-linked figures?
It’s not just about personnel. The language in the report reads more like a campaign manifesto than independent research. It posits a binary choice: pro-growth DA pragmatism versus ANC-driven “pro-poverty” redistribution. Framing that conveniently flatters one side. The VAT hike is portrayed as a turning point, but the IRR fails to fully acknowledge its own poll’s ±4% margin of error or the volatility of non-election-year sentiment. The sample size? Just 807 respondents. Significant, yes, but hardly a definitive indicator of national mood.
Moreover, the IRR explicitly states that “no turnout scenarios were applied.” This matters. Polls conducted in off-cycle years, without accounting for who’s likely to vote, often overstate the enthusiasm of the most vocal and politically active segments, like DA supporters riled up over tax policy.
And let’s not forget the obvious, that the DA is part of the GNU. So are the ANC, IFP, PA and others. The IRR praises the DA for opposing a policy enacted by a government it helped install. It’s a slick political two-step to reap the benefits of governance when convenient, and retreat to opposition posturing when it’s not.
This poll, then, is less a snapshot of South Africa’s political reality than a strategic narrative that the DA is rising, the ANC is falling, and the IRR just happens to be the first to break the story. But when the pollsters are friends of the party that benefits, we owe it to democracy to ask if this is research or is it a press release?
There’s no crime in being a partisan. But let’s not confuse a cheerleader with a referee. The IRR may present its poll as fact, but facts, too, are shaped by the hands that gather them.
For more political analysis and commentary in vernacular, join the conversation on Rabbie’s www.youtube.com/@Rabbie_Wrote