What the polls are saying about Spain’s upcoming national election

Spain is set to hold an early national election on July 23, after Prime Minister Pedro Sánchez of the Spanish Socialist Party (PSOE) brought it forward in the wake of very poor results in local and regional elections in late May. Mirroring similar trends elsewhere in Europe, Spain appears to be swinging to the right, with the far-right Vox party making significant gains.

According to our average of polls, whose methodology is explained at the end of this article, the conservative Popular Party (PP) would be the most voted party, hovering around 34% of the votes, followed by the PSOE (27%-28%), Vox (14%) and the leftist electoral platform Sumar (13%). But these figures keep shifting, as the graph below shows:

Contradicting my impression from last week, the PP now appears to be flat or even growing slightly to the detriment of Vox, which loses a few tenths of a percentage point.

The most important trend, however, is something else: the PSOE continues to make headway, as it has been doing since this snap election was called. The Socialists have gained two points since June, and what is more important to their interests: their step forward is not accompanied by a setback for Sumar, but for other parties.

This can be verified by looking at the evolution of the vote of the left bloc (PSOE and Sumar) and the right bloc (PP and Vox):

The sum of PP and Vox is still ahead (47% in votes), but PSOE and Sumar are recovering ground (41%). If at the beginning of June the advantage of the right was nine points, it has now been reduced to six or seven. The makeup of Spain’s next government depends on that figure, more than on any other, on whether it ends up being a +4 or +7 for the right.

Probable majority: PP and Vox

How do the above polls translate into parliamentary seats and government options? Below I have calculated the distribution of seats. I do so assuming that the national vote totals will be distributed by the different provinces in a similar way as they were in 2019 (see methodology). In this main scenario, the sum of PP and Vox is around 180 representatives and would still have a slim majority (the threshold is at 176).

In essence, the current polls see a PP and Vox government as the most likely option, although it is far from being a certainty. This is so for a simple reason, which I have already mentioned: both parties only have half a dozen seats of margin, and the polls are an approximate exercise. How approximate? I have analyzed dozens of elections since 1986 to measure this. The average error of the average poll has been around 1.9 points per party. That is, deviations of 3 or 4 points were common with some major parties, and the margin of error (at 95%) is close to 10 or 12 seats up or down from what was predicted.

The left’s options

The possibilities for the left are to either continue its advance in the remaining two weeks or to pull off a surprise on election night. As I said last week, the former seems plausible: Sánchez has made the rounds of radio and television stations and this could give him a boost, even if only by mobilizing his own electorate, which is very passive.

But how much does the left have to gain or surprise in order to change the most likely government? If the sum of PSOE and Sumar advances 1.5 points over the PP-Vox figures, the blocs would be 42% to 46% in votes, and the right’s majority would be up in the air, at a 50% probability, with the same chances of achieving it as of falling short. If the PP remains very close to the threshold of 176 seats, it could seek the support of a party other than Vox to get the next prime minister ratified by Congress, perhaps by convincing the small regional parties Coalición Canaria or Teruel Existe. However, if that sum is not enough, the alternative for the left would be to cobble together a majority similar to the one that made Sánchez the government leader four years ago, with the support of Sumar and practically all the other forces.

Averaging methodology. Our average takes into account dozens of polls to improve its accuracy, taking all those published by the media and collected in Wikipedia periodically. It is a weighted average to give different weight to each poll according to three factors: the pollster and its history, the size of its sample, and the date of the study. The last criterion - giving more weight to more recent polls - is the most important of all. In the average I have excluded polls by the public research institute CIS, which have been shown to be biased, and other pollsters with no track record. I have used similar methods since 2015.

Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition