# Andalusian elections: first real battle between polls and crypto markets

> Polls or crypto markets? Javier Callejo's analysis of how the Andalusian elections have become the crypto electoral test
> By Javier Callejo · Sociological Analysis · 2026-05-17

The Andalusian regional elections of 17 May 2026 have arrived with a rarity unusual in recent Spanish politics: a general coincidence between opinion polls and crypto betting markets on the winner, but not so much on the exact scale of the victory. Above all because of the abundant and rapid swings in the betting markets' prediction of the likelihood that the PP will reach an absolute majority.

Both survey-based polls and prediction markets take the victory of Juanma Moreno's Partido Popular as certain. However, the great unknown of the campaign — and of recent weeks — has been whether the PP would or would not revalidate its 2022 absolute majority. That is where the most interesting differences between the two electoral forecasting systems appear. Revalidating the PP's absolute majority is, in fact, the news focus of these elections.

## The PP and the 55 seats

The political heart of the event is whether the PP manages to exceed 55 seats. That is where the Andalusian political system is at stake and, by extension, the Spanish one. But it is also the point where polling institutes and prediction market bettors are at stake.

Beyond the money wagered, there is a new opportunity to observe which electoral forecasting system is the most accurate, with the systems of so-called predictive intelligence gradually being added. In any case, we are thus facing this new "electoral classic", the one that pits survey-based polls against betting markets. An "electoral classic" in Andalusian key. Sporting language prevails when it comes to comparing systems to answer the question of what people will do before people themselves know it.

**[The crypto market tops $150,000 on the Andalusian elections and doubts Moreno's majority](https://observatorioblockchain.com/en/blockchain/150-000-dolares-en-cripto-en-juego-por-la-mayoria-de-moreno/)**

At the close of this edition, the volume traded in crypto markets on the Andalusian elections is already around $300,000 according to Polymarket records. A relatively small volume compared with major international elections; but it must be taken as a relevant sign of the growth of these markets in our country. It is hard to think that citizens of the world far from this Spanish region would participate, with their money, in this predictive market.

## The bettors

[A prediction market](https://polymarket.com/event/president-of-andalusia-after-election) relatively small for the Andalusian elections. If we compare it with, for example, other electoral contests of great international scope. Thus, the US presidential election on Polymarket moved billions of dollars; That low participation has an important consequence: prices and probabilities can be more volatile and less statistically robust than in giant markets.

A handful of bettors with relatively modest amounts can move the implied probabilities quite a bit. Beyond this call for methodological caution in interpreting the results and, above all, their swings, what interests us here is the comparison between survey-based polls and prediction markets. Continuing with sporting language, at this minute of the match the polling approach gives relevant probabilities to a result with an absolute majority for the PP.

The probabilities given by prediction markets keep changing, with a trend observed in recent hours toward giving that absolute majority, when at the close of the electoral campaign they denied it. That is, if at the end of today's election day the PP were to achieve an absolute majority, both prediction systems could show themselves as winners. Like the very life of elections, where all the competing political parties seem to have won at the end of it.

## The consensus of the polls: a clear PP victory

Throughout the campaign, the main polling institutes agreed on several fundamental points. Points such as that the PP would remain the hegemonic force in Andalusia or that the PSOE would obtain one of the worst results in the history of this region. The main points remain within two-party logic. The common denominator is the idea that Juanma Moreno would win clearly, although possibly — and this "possibly" is the important thing — depending on Vox to govern. We thus move from party logic to bloc logic.

And it is that the key points of these Andalusian elections fall within the opposition between a right/left bloc logic and a two-party logic (PP/PSOE), with the estimate of Vox's results being the crux of the matter. In this regard, there is also some agreement among polling institutes regarding baseline flows: Vox would consolidate a relevant position, although very far from disputing the leadership of the right.

## Alternative left

But the degree of that relevant position remains open. Very open. For its part, the estimates for the so-called alternative left, fragmented between Por Andalucía and Adelante Andalucía, pose a scenario in which it would struggle to achieve a result that would allow it to play a relevant role in the political period that opens after these elections. In short, the most open question is whether Vox would obtain enough votes for any of the following alternatives: a) to prevent the PP's absolute majority and force it into negotiation; b) even if the PP's absolute majority were not prevented, its electoral weight, especially against the PSOE, were especially important.

The aggregate poll results published up to the legal close of polling put the PP at around 43% of the vote and between 54 and 57 seats, right at the threshold of the absolute majority set at 55 deputies. The statistical model of 40dB. for El País and la SER calculated that the PP would reach an absolute majority in approximately 57% of simulations, a relatively high probability, although very far from mathematical certainty.

## Betting markets: more cautious than the polls

Betting and prediction markets, especially international platforms such as Poymarket, showed over the past week a somewhat different and, above all, changing reading. While the polls tended to place the PP on the edge of an absolute majority, bettors initially seemed more skeptical about the possibility of Moreno exceeding 55 seats.

In the final days of the campaign, many markets gave higher probabilities to the scenario of a broad PP victory, but insufficient to govern alone. Apparently, bettors interpreted that a small final growth of Vox could cost the PP the last decisive seats. But such results have changed in recent hours, showing a higher probability that Moreno will achieve an absolute majority.

## What the polls ask

In other words: while polls ask voters what they plan to do, markets bet on what they believe will actually end up happening. Two different methods for something as difficult as measuring the future. Now, far from the noise of sporting language, speaking of a clash between prediction methods greatly oversimplifies things. Even more so if one takes into account that, when it comes to election results, a good part of prediction markets' decisions rely on the analysis of the results that the various public opinion institutes offer through the media.

That is, prediction markets act as an expert system with respect to electoral polls. Prediction markets have a meta position — above in the cognitive, organizational and decision-making process — with respect to electoral polls based on opinion surveys. Perhaps for this reason, and knowing the latest published forecasts from polling institutes, they have leaned more toward taking Moreno's absolute majority as the result.

## Two different methods for measuring the future

The difference between polls and betting markets is not only technical; it also reflects two different ways of understanding political prediction. Polls are based on representative population samples and electoral estimation models. Their goal is to photograph voting intention at a given moment. The key question is: "And in the next elections to the Parliament of Andalusia on 17 May, which party or coalition would you vote for?" (question 12 in the CIS pre-election study, published on 24 April 2026).

The response of the people interviewed, like the interaction between informant and interviewer, may be mediated by multiple factors that are not easily controllable, such as: electoral context, concealment of vote, social desirability, etc. For this reason, they require the work and experience of professionals to make the estimates, which tend to be delegitimized under the term "cooking".

Now, cooking, as in our home, is necessary. And what many of the participants in electoral prediction markets do is exactly that: do their own cooking-estimation based on the poll results.  
Prediction markets work differently from electoral polls based on opinion surveys. They do not ask what people will do, but rather operate on the money people are willing to risk on a given result. In this case, for example, whether the PP will obtain an absolute majority or the number of seats each political party will win.

## The money they wager

To decide on the money they wager, on the position in play that they buy at the price set by the prediction market, bettors analyze the results of pre-election polls, take into account previous electoral behavior in similar contests and the differences between final results and those forecast by the polls themselves. They act as an expert system. As a second-order observer, as professors of social systems theory would say.

Of course, it is a decision that can also take into account changes in the political climate, special public behavior of candidates appearing in the media, rumors or turnout trends and, from there, speculate on the potential tactical movements of the electorate.

Therefore, it is not an equal clash between two competing systems, however much sporting language insists. It can be said that initially the prediction market is at a higher logical level. It has a meta position with respect to polls, to put it plainly and directly.

## More mature prediction markets

For this reason, from the point of view of an external observer, of a kind of "third observer" — an observer who observes how the second observer (prediction markets) observes the first observer (polls) — they would be more reliable, provided that the volume of the prediction market, for the specific bet, is large enough. Otherwise, the prediction resulting from these markets may be affected by speculative movements of few participants and show significant swings.

In more mature prediction markets, which are those to come in the future, the relationship between polls and these markets may become more complex. One might think that those who respond to the poll have bet in one of these markets or that, at least, they are aware of the results being produced and the quotes of the various alternatives.

Their response may then be mediated by the availability of such information. In that case, it would be the polls that would act as a second-order observer, as an expert observer. We would find ourselves in a rather infernal circle and, at the same time, a major challenge for electoral prediction professionals.

## A political and professional laboratory for Spain

The Andalusian elections of 2026 have once again turned Andalusia into an advanced laboratory of Spanish politics. But also, and taking into account the different predictions of pre-election polls based on opinion surveys and prediction markets, a laboratory for prediction and electoral analysis professionals.

The polls have been showing for weeks a community consolidated as a PP stronghold, where the party is relatively close to revalidating the absolute majority it currently holds. Betting markets, by contrast, have been less stable, between absolute majority no, absolute majority yes. Election night will tell who came closer to being right: traditional polling or the collective — and financial — intelligence of prediction markets.
