Colombia Election Notes - 30 April 2026

The polls say Cepeda is the frontrunner. I'm contrarian in viewing Valencia as the likely, but not certain, winner.

Colombia has seen lots of new polls in the past week. Quick summary:

Atlas:

• First Round: Cepeda 37, de la Espriella 29¹, Valencia 21

• 2nd Round: Valencia: 49, Cepeda 41

• 2nd Round: de la Espriella 48, Cepeda 42

RCN/GAD3:

• First Round: Cepeda 36, de la Espriella 21, Valencia 13

• 2nd Round: Cepeda 44, Valencia 37

• 2nd Round: Cepeda 46, de la Espriella 35

Invamer:

• First Round: Cepeda 44, de la Espriella 22, Valencia 20

• 2nd Round: Cepeda 51, Valencia 47

• 2nd Round: Cepeda 55, de la Espriella 43

CELAG

• First Round: Cepeda 42, Valencia 20, de la Espriella 20

• Cepeda 48, Valencia 39

• Cepeda 49, de la Espriella 36

I’m still waiting on Guarumo and CNC to publish their late April polls.

Takeaways relying on just the polling

Cepeda remains the clear first-round frontrunner. Seems obvious.

The Petro opposition is divided. The divide isn’t just between Valencia and de la Espriella, but includes a mix of other candidates making up 10-20% of the vote plus undecided voters who are likely waiting to determine how they will best cast a “voto util” against the ruling party.

Valencia is a stronger second round candidate than de la Espriella. It’s only a few points of advantage, but in a close election, opposition-leaning voters will take that edge.

The center has collapsed. Claudia Lopez and Sergio Fajardo are both still hanging on in fourth and fifth place in most of these polls, but their numbers have both dropped since the primaries.

Despite these topline numbers, the polls, while accurate representations of public opinion now, don’t reflect how this election will turn out in the coming weeks. It’s bad practice to pick polls apart in terms of methodology, sample, and demographics to try to hack my way toward a preferred outcome. It’s generally best to take the polling average as is. However, the late-April polls are missing structural shifts that will reshape the race by late May. My comments here explain why prediction markets are not as certain about giving Cepeda full frontrunner status.

Prediction markets have the race around Cepeda 40% likely to win, Valencia 35% likely to win, and de la Espriella at 30% likely to win. Reminder that those numbers are not polls, but rather probability odds based on actual money people are investing/betting. That’s a close, almost three-way race. It’s true that there has been some market manipulation in prediction markets over the past months, but as this election gets closer and more money flows into the market, manipulation becomes more difficult.

My contrarian call

My prediction remains that Paloma Valencia is the favorite to win this election. My current prediction on the odds for the next president is about Valencia 55%, Cepeda 35%, other candidates including de la Espriella <10%.

I decided to send the above comments about the polls to everyone today. Additional analysis of my election prediction is available to paying subscribers below.

Factors that make me believe Valencia will outperform the current polls.