The new spatial model of voting in Spain
31 December 2014

Kiko Llaneras published last week a very interesting collection of figures describing the new political arena in Spain. In my opinion, the most remarkable message from his post is that, in spite of Podemos’ effort to become a partido transversal, a party that operates outside the narrow scope of ideological competition, if one takes the political surveys from the CIS what we see is a party that is located somewhere between IU and PSOE, fighting in an already crowded space. It is (not so) surprising once you think that the common perception is that Podemos is capturing disatisfied and disaffected voters, a condition that does not need be tied to a particular ideology. However, it seems that things behave as in the spatial model that we all know and love.

The interesting figure is shown below and depicts the distribution of ideal points for voters that indicated closeness (simpatía) to each of the main political parties:


The extent to which the ideogical location is sincere or the result of some ex post adjustment in the perception of the political space is an essential question, but the data needed to tackle this question is not available, as far as I know.

Unfortunately, the original post remained in the descriptive sphere and simply summarized the raw data. If it is true that the spatial model works, we can (and should) go further. At the end of the day, all we need to pose is a model1 in which the probability of voting for a given party declines with the perceived distance2 in order to estimate the parameters governing the behavior of voters. If we truly believe the spatial model, then we must conclude that political competition in Spain looks as follows:


For each ideological location, the figure plots the predicted probability with which a voter with that ideal point supports each party, assuming there is full turnout and that the four parties represent the complete menu of options.

The figure shows Podemos competing with the PSOE in all the political space in which the PSOE traditionally captured voters (left and center-left), and essentially controlling the leftmost side of the ideological spectrum —most likely for purely non-spatial reasons. If the figure is true, any small movement of Podemos towards the center will put PSOE in a dangerous situation, in which traditional PSOE voters will end up deciding which party to support through a coin flip. But again, that scenario will only be true if the movement actually happens —the party has to consolidate its greyfication—, if the movement is credible —voters should believe the party leaders wants to implement its manifesto— and if it is able to maintain the valence advantage —the image of outsider with respect to the big, traditional parties— that has brought it here. And about that, the spatial model has little to say.3

  1. Contrary to the widely held belief, data is useless without a theory. 

  2. There are a number of assumptions that I should have refined that I simply accepted in this case for the sake of simplicity. 

  3. Replication code avaliable here