The North American electoral betting market generated around US$3.5 billion in Trump’s election last year. Anticipating the future has always been a human obsession. From divination to astrology, from prophecies to scientific projections, the journey of millennia has summoned our insatiable curiosity to prevent tragedies, win wars and plan the future. Since the ballot box became a substitute for bloody disputes for power, they have attracted the same fascination. This explains the volume of analysis on the 2026 presidential succession.
From now on, the number of surveys will grow, driven by different interests, attentive to fluctuations in popular mood, even before the full list of competitors is known. The institutes test multiple lists, while potential candidates and party leaders insinuate or deny claims. A new fact, the manifestations of candidates for vice-president are multiplying, remembering the time when they competed in their own category, and not attached to the incumbents.
Once again, the polls will act as a safe handrail guiding the press, campaigns and citizens through the fog in which information and disinformation mix. What is certain is that the dance of numbers will not stop until the big day. Still, the temptation to glimpse the outcome remains irresistible.
Based on the 13 competitions held since 1945 — 4 in the pre-64 democratic cycle and 9 in the — it is possible to identify patterns with reasonable consistency. The first concerns the configuration of the confrontation. In around 70% of cases, three names reached double digits, forming triangular scenarios. This seems to be the most likely format in 2026, as the 2022 bipolar duel will not be repeated, and there will be more than one candidate on the right.
Another issue: the possibility of winning in a single round seems reduced, just 23%. it never exceeded half of the valid votes, although it came close in 2006 and 2022. Only two did so: Eurico Dutra (record holder with 55%) and FHC, in his two consecutive victories.
As for the historical average of winners in the first round, the rate is around 47%. Excluding 1989, the most fragmented, it rises to 48%. With variations of less than two points, this is the modal value of the results. Five of the six electoral days of this century confirm the pattern. Therefore, whoever prevails in 2026 should be close to these marks.
Finally, the most complex question: what are Lula’s prospects? Necessarily restricting ourselves to the New Republic, it is observed that the current continuity rate —75%— is a limited parameter, as it rests on just four cases. Nor is it appropriate to compare with foreign experiences given the significant differences. It is better to observe voting intentions a year before the election — 6 of the 9 times the initial leader ended up victorious. And combine them with government approval data, expectations about the economy and market projections of key variables — GDP, inflation and unemployment — for next year.
Considering these factors, with relatively benign macroeconomic signals, our parsimonious model with Bayesian logistic regression, although sensitive to the short historical series, indicates a predictive and updateable projection of 72% chance of reelection for the incumbent. Which will adjust as new empirical evidence emerges. If the economic situation slips and disruptive political events occur simultaneously, the balance tips towards the opposition. Likewise, stability and growth reinforce the president’s position.
Remembering that probabilities are not predictions. It’s always good to hear Shakespeare’s message: “Never call impossible what seems improbable to you.” This was didactically illustrated in Trump’s first presidential run, when the renowned American prognostics site FiveThirtyEight gave him chances of just 20% the day before.
