Democracy Is Becoming a Betting Market

By Chenayde Abrahams and Kavisha Pillay

South Africans are preparing to vote for the people who will govern their cities, fix their roads, keep the taps running and restore some semblance of dignity to communities that have endured years of political dysfunction. Meanwhile, bettors from the country and around the world are putting money on who will emerge victorious from the political chaos.

On prediction markets such as Polymarket, the future leadership of Johannesburg, Tshwane, Ekurhuleni, eThekwini and Nelson Mandela Bay has become a financial proposition. People can buy and sell contracts predicting who will become mayor, with money to be made from getting the outcome right.

The absurdity of this is clear in Johannesburg, where residents routinely contend with water outages, crumbling infrastructure and unstable coalition governments, but the question of who will lead the city has become an opportunity for financial speculation. The very instability that makes life so difficult for residents can now create further opportunities for people to profit.

The market is growing

Public listings reviewed by the Campaign on Digital Ethics (CODE) ahead of South Africa's local government elections showed approximately US$394,000 in cumulative trading activity across eight metropolitan mayoral markets on Polymarket. Other platforms, including Kalshi and the South African betting exchange Betmatch, have also offered election-related contracts.

Prediction markets allow participants to trade contracts linked to future events. If a contract predicting a particular outcome is trading at 60 cents, that price is commonly interpreted as a 60% implied probability of the event occurring. The contract pays out if the prediction proves correct.

The price, however, reflects the behaviour of people participating in that market. It tells us nothing directly about how representative those participants are of the electorate, what information they possess or whether a handful of wealthy traders account for much of the activity.

That distinction is becoming increasingly important as prediction markets move from the margins of internet culture into mainstream political discourse.

The markets got Trump right

The 2024 United States presidential election was a defining moment for the prediction-market industry.

As opinion polls suggested an exceptionally close contest between Donald Trump and Kamala Harris, platforms such as Polymarket and Kalshi assigned Trump a higher probability of winning.

On the eve of the election, Polymarket priced Trump's chances at approximately 57%, while Kalshi put them at around 51%. Trump subsequently won.

For advocates of prediction markets, the result was a powerful advertisement. Here, they argued, was evidence that markets could aggregate information and anticipate political outcomes more effectively than conventional polling.

This is a serious argument worth considering. People who risk money on their predictions have an incentive to pay attention, process information and correct mistaken prices. A functioning market may incorporate developments more quickly than a survey conducted several days earlier.

The industry's triumphalism, however, deserves scrutiny.

A market that assigns one candidate a 57% chance of victory is expressing uncertainty. It is not making the same claim as a poll that measures voting intentions. Furthermore, correctly identifying a winner in one election does not establish that market prices are consistently well calibrated.

There were also questions about who was moving the market. During the 2024 campaign, a French trader was identified as controlling several accounts responsible for millions of dollars in bets favouring Trump. Polymarket said it found no evidence of manipulation, but the episode demonstrated how concentrated financial activity could influence prices that were being widely circulated as political forecasts.

The concern is not confined to whether someone deliberately manipulates a market. Even sincere traders can introduce bias. Participants may be disproportionately drawn from particular countries, income groups, ideological communities or online cultures. Money creates an incentive to be correct, but it does not magically eliminate the assumptions and prejudices people bring with them.

Then came Brazil’s 2026 election, which exposed the limits of market certainty

In Brazil's presidential runoff on 25 October 2026, Polymarket told a very different story.

In the final days before voting, its presidential winner market heavily favoured Flávio Bolsonaro over incumbent Luiz Inácio Lula da Silva. On the eve of the vote, Bolsonaro's contract was trading at an implied probability of approximately 80%, while Lula's stood at roughly 20%.

Lula ultimately won, securing 51.79% of votes against Bolsonaro's 48.21%.

The outcome was a striking reminder that a market price, however confidently presented, remains an estimate produced by a particular group of traders.

Analysis of the Brazilian market has raised questions about the information available to participants and the composition of the trading population. Brazil had restricted access to prediction-market platforms, potentially limiting participation by people with local political knowledge. Some analysts also suggested that traders placed excessive weight on Bolsonaro's first-round performance and failed to account adequately for second-round turnout and the transfer of votes from eliminated candidates.

But the Brazilian example should trouble anyone tempted to treat prediction markets as a superior form of democratic intelligence. A contract price is shaped by those willing and able to trade. It may reflect expertise, but it may also reflect incomplete information, concentrated wealth, limited participation and collective misjudgement. The fact that people are prepared to wager millions on a political outcome does not make their expectations an authoritative account of political reality.

South Africa's coalition politics create particular risks

In South Africa, the problem becomes even more complicated because voters do not directly elect metropolitan mayors. We elect municipal councillors, who subsequently choose mayors through council votes and, frequently, coalition negotiations.

A bet on Johannesburg's next mayor is therefore a wager on a political process that extends beyond election day. It involves expectations about party negotiations, possible alliances, leadership decisions and the distribution of political power.

Much of that information may initially be known only to a small circle of political actors. What happens when someone involved in coalition negotiations knows that an agreement is about to be concluded, or that a candidate has secured decisive support, before the information becomes public?

The possibility of profiting from privileged political information creates an obvious conflict of interest.

The United Kingdom's 2024 election-date betting scandal illustrated how people with access to confidential political information may have opportunities unavailable to the public. South Africa should not wait for a comparable scandal before establishing appropriate safeguards.

The stakes are particularly high in municipalities where coalition instability has already weakened public confidence in political institutions.

We should also ask whether commercial betting creates incentives that sit uneasily alongside public responsibilities. Political parties and elected representatives are supposed to make decisions in the public interest, subject to democratic accountability. The emergence of financial markets around those decisions creates new questions about disclosure, conflicts of interest and the possibility of information being exploited for private gain.

South Africa is already confronting the social and economic consequences of an expanding gambling industry. National Gambling Board figures show approximately R1.5 trillion in gambling turnover and R74.9 billion in gross gambling revenue during the 2024/25 financial year.

These figures illustrate the enormous scale of gambling activity in a country marked by unemployment, indebtedness and profound economic inequality.

Prediction-market companies frequently distinguish their products from conventional gambling, describing them as financial exchanges or information markets.

That distinction may be relevant to their design and regulation, but it does not settle the public-interest question.

South Africa needs clear rules on political event contracts, particularly those offered by offshore platforms. Regulators should establish which activities fall within existing gambling legislation, what licensing requirements apply and how consumers and democratic institutions are protected.

There should also be explicit safeguards against trading on confidential electoral or political information, transparency around large or concentrated positions, and meaningful scrutiny of markets susceptible to manipulation.

Newsrooms have an equally important responsibility. If betting prices are reported, audiences should be told what they measure, how much trading has occurred and why they are not equivalent to polling data.

Above all, we need to resist the growing assumption that every dimension of public life should be turned into a market.

At CODE, our concern is with the increasingly powerful commercial systems that shape how people encounter, understand and participate in democracy. Prediction markets are becoming part of that system. Our democracy is already paying a high price for the concentration of political and economic power. We should be very careful about allowing another industry to profit from the uncertainty of its future.