The most surprising result of the 2015 general election in Croatia was the success of the recently local party MOST, that appeared to have broken the duopoly of the two establishment center-left and center-right parties, the SDP and HDZ. MOST was founded in a small coastal Croatian city to compete in its local elections in 2013. After winning those elections they spread out on a national basis attracting other local, independent, uncorrupt mayors, and in just 10 months since the beginning of 2015 they turned themselves from a local movement to a kingmaker party winning a total of 19 seats (out of 151), making it impossible for both establishment parties to form a government without them.
In our model we were unable to correctly predict the success of MOST. In some districts the probability of their success (winning 2 or 3 seats out of 14) had a probability as low as one in 10,000. It was a typical fat tail (black swan) event, that went under our radars. In our final forecast we predicted them a total of 8 seats. In the elections they managed to more than double that number.
The story of third party politics
Before we even started making predictions we knew that predicting the outcome of the first two establishment parties, HDZ and SDP, wasn’t going to be that difficult. We had a lot of data on their past electoral results, and a number of polls tracking their trends on a monthly basis. We knew that the difficult part would be to forecast the so-called third options. Particularly since in these elections there were a total of 10 completely new parties being formed since the previous elections (some of them from washed-up politicians, some from completely new players on the national political arena – such as MOST). We had no data on MOST whatsoever, but we had a way to predict the success of the “third option” parties. To understand how we should briefly describe some of the specifics of the Croatian electoral cycles so far.
In the past 15 years a steady number of Croatian voters have always been searching for the so-called third party – a party that’s neither the center-right HDZ, nor the center-left SDP (both of which are sharing the blame for economic failures, corruption, emigration, etc.). In this search for a viable alternative a number of parties and candidates have shown up on different elections. However the specific electoral system in Croatia (the D’Hondt method is used for calculating votes into seats with a threshold of 5%), has been unkind to third options. Usually between 300 and 600 thousand votes are being wasted each elections by going to parties which fail to pass the 5% threshold (one should note that Croatia has over 100 registered political parties). It is thus easy for the two establishment parties to maintain their dominance, basically turning the electoral system from a proportional to a majoritarian one (albeit with coalitions with minority parties necessary every time in order to form a government). Across the past 4 presidential and 4 general elections, in almost each of them there was a party or a candidate that arose as a third option, with varying degrees of success, however each time important for government coalition formation, or at least gaining significant media attention after the event.
There are two kinds of third party options: (1) parties that take votes away from one of the two dominant parties, and (2) parties that take votes away from both parties and/or attract new voters. In the following set of correlation graphs the x-axis shows the vote share for HDZ (blue) and SDP (red), while the y-axis shows each of the third option parties for different elections, for different electoral districts on a voting station level.
The first two graphs show the stipulated relationship for ORAH and the Labor party (Laburisti) for two different elections. In both cases these two parties took votes away from the center-left SDP (as both of these parties positioned themselves on the left political spectrum). The other two graphs show a clear negative correlation between the third option parties and both HDZ and SDP. The reason is that both of these parties (HDSSB and IDS) are primarily local parties operating only within one or two electoral districts, and are focused on utilizing their support from those regions.
What we managed to find out by analyzing each individual success of different “third option” parties across all these elections is that they always gained significant support in roughly the same geographic areas. This is generally the north-western part of the country, the capital Zagreb (to some extent), parts of Slavonia, the coast of Dalmatia and the Istria region (see the map below). Note that electoral success for these options is defined as anything between 10% and 20% of the vote share in any electoral district. There has never been a district in which a third party overcame any of the two major parties.
The success of Most
Comparing all these options together we could roughly estimate the total success of all “third option” parties for each district. We assumed that the usual voters for the third option parties would split their votes across the ten (or at least five) newly formed parties, with one of them coming on top – MOST. However, we predicted a total of around 18 to 20 seats for third option parties to be evenly distributed amongst at least five parties. This didn’t happen since MOST alone took almost all of these votes, leaving most of the other parties humiliated (one party, which at one point boasted a nationwide support of 19%, failed to even pass the 5% threshold).
To account for this unexpected success we needed to see where the new party MOST drew most of its support from. It was, as expected, from roughly the same areas where third option parties usually did well in the past. Furthermore, their success had a high correlation with the success of some previous presidential candidates that were running as independent, third option candidates – most notably Boris Miksic, the candidate that came third in the 2005 elections, with 20% of the votes (the second placed candidate got 21%), or roughly 400 thousand voters. MOST received about 300 thousand, but their geographical distribution of voters was very similar as the one for Miksic in 2005.
This implies that Most was able to galvanize the support of mostly swing domestic voters which are in constant search for something new. They failed to attract new voters, since turnout numbers were even slightly lower for this election than for the last one.
To see this in more detail we needed to examine whose voters MOST took over, with respect to both previous results of other third option parties, and the socio-economic data (unemployment, education, income). They mostly emulated the success of former third option parties (such as the two presidential candidates Sincic – yellow, and Miksic – purple), which was as expected from our earlier analyses.
Comparing their result to the 2011 elections it seems that MOST took mostly SDP voters (the SDP-led coalition achieved a great result at those elections, more than a million voters, while they now dropped to 740 thousand). However looking at the results of the 2014 EU parliament elections, and the 2015 presidential elections, it seemed that MOST took the voters mostly from HDZ (see graphs). Both of these theories are true. We are in both cases talking about one group of swing voters who most of all like to punish incumbent parties, dissatisfied with their in-office performance. These swing voters punished the HDZ government on the 2011 elections and kept punishing the SDP government on the EU and Presidential elections in 2014 and 2015. However, during 2015 MOST gained momentum and was able to attract a large majority of these swing voters for themselves. HDZ was counting on them to punish the incumbents and swing back to them, but they instead chose Most. They did punish the incumbents, but not entirely as expected.