How to solve the problem of registered voters not casting their ballots

South Africa is, once again, months away from a local government election. On 4 November, more than 28.7 million registered voters — a record — will have the opportunity to walk into a voting station and shape who runs their municipality. It is the single largest civic exercise this country undertakes outside of a national election, and it will, once again, cost the taxpayer, the Electoral Commission of SA (IEC), and every contesting party enormous sums of money to stage.

Are we spending that money on the right problem?

For a decade, South Africa’s political establishment has treated voter registration as the metric that matters — the number the IEC leads with in every press briefing, the number parties mobilise around, the number that makes headlines. But registration was never our problem. Our problem is what happens after people register. And the data, going back a full election cycle, make that painfully clear.

Read that table again. Nationally, the voters’ roll barely moved — a decline of less than half a percent over five years. But every one of our nine provinces lost between 11% and 25% of its valid votes. Gauteng and Mpumalanga each lost roughly a quarter of their 2016 voters. Nationally, 3,216,029 fewer South Africans cast a valid vote in 2021 than in 2016 — despite an almost identical pool of people eligible to do so.

This was not a registration failure. The IEC’s machinery held the roll steady. This was a turnout collapse: national participation fell from 56.80% of registered voters casting a valid vote in 2016, to just 44.81% in 2021 — a 12-percentage-point drop in a single electoral cycle, the lowest local government turnout in the history of our democracy.

2026: Registration surge and an open question

This cycle looks different, at least on the metric we keep measuring. As of 30 July 2026 — one day before the IEC’s final in-person registration weekend closed — the national roll stood at 28,738,902, a 9.67% increase on 2021, the sharpest growth in over a decade.

Nearly 2.5 million new names have been added to the roll — with Limpopo and North West leading the growth, KwaZulu-Natal and the Eastern Cape leading in raw registration-day volumes. On paper, this is a genuine success for the IEC’s online portal and its youth-focused “Get Up, Show Up, Vote” campaign.

But we do not know how many of those 28.7 million will actually vote. That number only exists after 4 November. If the 2016–2021 pattern repeats — a healthy roll masking a collapsing turnout — this registration surge will prove to have been an expensive exercise in optics rather than democracy. If it doesn’t repeat, and this becomes the election where registration finally converts into participation, it will be the first time in a decade that it has. Either way, in a matter of days after polling, we will know.

What other democracies have worked out

South Africa is not the first country to discover that a healthy voters’ roll and a healthy democracy are not the same thing. At least 21 countries have concluded that voluntary turnout alone cannot be trusted to deliver a representative outcome, and have made voting compulsory. Belgium has enforced it since 1893. Australia’s modest fine for not voting keeps turnout above 90%. Argentina, Brazil, Uruguay and most of Latin America enforce voting through fines and administrative friction — in Brazil, an unpaid voting fine can delay a passport application or a bank loan.

The result, unsurprisingly, is turnout in the 80–95% range — a figure South African democracy has never come close to at local government level.

But compulsion alone is a blunt and, frankly, illiberal instrument if it stops there. Forcing a citizen to the polling booth without giving them an honest way to say “none of these candidates deserves my vote” manufactures the appearance of legitimacy while masking the same disengagement it claims to solve. This is where the more sophisticated democracies have gone further, building a “None of the above” (Nota) option directly into the ballot:

• India introduced Nota in 2013 after its Supreme Court ruled that the right to reject a candidate is inseparable from the right to vote. Nota votes are counted and published, but do not change the outcome — a purely symbolic, if powerful, protest.

• Colombia goes further still: a blank vote is constitutionally recognised as “a political expression of dissent”, and if rejection wins a race, fresh elections are called with an entirely new slate of candidates — a real, structural consequence for a party field voters reject outright.

• Brazil combines both ideas — compulsory turnout and a formally tallied blank/null vote, separate from spoiled or accidental ballots. In its 2022 general election, roughly 5.2% of all votes cast were blank or null: a measurable, undeniable signal of dissatisfaction, inside a system that still required the whole electorate to show up.

That combination is the model worth South Africa’s serious attention. It solves the exact problem our data exposes. Right now, the 3.2 million South Africans who declined to vote in 2021 are invisible in our national conversation — folded into a vague, unmeasured category we call “apathy”. A compulsory-vote-with-Nota system would convert that invisible disengagement into a hard, published, undeniable number — one that every underperforming councillor, mayor and party leader would have to reckon with by name, ward and municipality.

The closing argument

South Africa now has more registered voters than at any point in its democratic history — more than 28.75 million names on the roll, a 9.67% jump in five years that the IEC, political parties and civil society worked hard to deliver.

But registration has never been the country’s problem. Between 2016 and 2021, the roll barely moved (-0.49%) while 3.2 million fewer citizens actually cast a valid vote — a 12-point collapse in real participation that no registration drive, however well-funded, can fix on its own. Whether the class of 2026 turns out on 4 November, or simply joins the growing ranks of registered-but-absent citizens, will be the real verdict on whether this investment paid off.

It is time our legislators seriously consider electoral models already proven elsewhere: a system of compulsory participation, paired — as Brazil, Colombia and India have shown, is both possible and fair — with an explicit, counted “none of the above” option on the ballot.

Such a system would not force any citizen to endorse a party they do not trust; it would simply convert South Africa’s silent, invisible disengagement into a counted, undeniable political signal — one that would cost every underperforming party and municipality far more than a spoiled ballot ever could. If upholding democracy is worth the billions the IEC spends running these elections, it is worth this Parliament’s serious consideration too. DM