GSAC Analytics

Preliminary analysis of the course and results of the parliamentary elections in Georgia

Preliminary analysis of the course and results of the parliamentary elections in Georgia

Well, I finally finished the preliminary assessment of the election results. There is still a lot of work and it must be done as quickly as possible.

The main research tool I used was the Kisling – Shpilkin method. I will not spread much about him, he is quite famous and easily googled by the name of his author Sergey Shpilkin. The fundamental hypothesis of the method is the assumption that the distribution of votes is Gaussian (normal) in different polling stations. The hypothesis is quite strong, so the method is often criticized. Nevertheless, the tool is very productive, especially if the falsifications are large-scale. In our case, the falsifications are not very large-scale (although they are decisive in the outcome of the elections), therefore, it should be treated with caution and used for a rough assessment or as an assistant in the search for questionable sites. Moreover, the opposition will have to sue in each case based on actual violations of the law, and unfortunately there is no legal norm “non-Gaussian distribution”.

Here I will attach static pictures with trend lines. But for those who want to help the opposition or study the situation in more detail, I have prepared interactive scatter charts on our GSAC-politics portal (link in the commentary). Moreover, for those who want to spend a lot of their time and help a lot with the search for falsifications, I have prepared an interactive map, but, unfortunately, I do not have enough free mapbox limits to open access to everyone. Contact us and I will open a personal account for you with access to the card (for 40-50 people, the limits should be enough). So, statistically falsifications are visible. I have divided Georgia into three large groups of areas in order to neutralize the influence of geography and regional characteristics.

Tbilisi’s schedule looks like a healthy person’s schedule. An increase in turnout lowers the power outcome. This is a normal situation when people are more active in voting with the intention of changing the government. That is why “Georgian Dream” during the pre-election period chose a strategy of moderate intimidation with covid (it tried not to lose face, but to cut off unnecessary people who are afraid to go to the polling station). Nevertheless, the trend line tail, which is pulled up, is alarming. Basically, he is pulled by the areas that are located in remote areas of Tbilisi (Okrokhana, Kojori …) Traces of manipulation are clearly visible on the graph of Eastern Georgia. It is not the tail that is pulled up, but the entire trend line. Explicit outlines are also visible (points far from the central heap). A similar situation is observed on the chart for Western Georgia.

Now about the main thing. What impact does all this have? I would agree with both independent vote counting (45%) and Ipsos Exit floor (42%). If you “cut off” the tails, then the result is obtained precisely in this range.

And of course, you cannot do without reservations and comments.

1.This method does not track the possibility of a massive transport of voters to all polling stations. Although we actually know that the Georgian Dream system is based on this. Unfortunately, this has become so commonplace that Western partners have not paid attention to it for a long time, and we affectionately call it an administrative resource.

2. Attribution of votes in a precinct with an initially low turnout “throws” a point into the central pile, and it also disappears from the method’s radars. It is better to use the map to find such points. For geographically close areas, the Gaussian hypothesis is the most accurate.

3.In my opinion, the confidence of the authorities in a successful result played a cruel joke with them, so I had to hurry to draw the result. That is why, this time, “Zhvania’s ears” are sticking out of a huge number of final protocols. It remains to find them all. But there is also good news (if it seems good to someone?). Unlike Russia, they painted, although systematically, but not centrally, i.e. without ordering the result. Therefore, “Churov’s saw” is not visible (also can be googled).

4. As I said, this is only a preliminary analysis, which includes an exclusively proportional part of the elections. A much more important analysis needs to be done by majoritarian districts. In some of them the advantage is so small that if the results of just a few precincts are canceled, the authorities lose the candidate’s victory in the first round.

5. Now about the victims. In the regions, as expected, votes were stolen from the UNM. But Tbilisi has become such a strange exception that it is worth paying attention to. With the apparent “normality” of the distribution of the results of the “Georgian Dream”, the “Strategy of Agmashenebeli” party was robbed. In another way, it is extremely difficult to explain the graph of the results of this party in Tbilisi.

6. Well, a few words about the political situation. This time it is very different from what it was before. Firstly, they stole the victory of the opposition, and not extra interest to satisfy their ego. Secondly, having understood the situation, it was too late to draw. There are many cases when the result changed on the way from the precinct to the Central Election Commission. As a result, in addition to the usual almost unprovable administrative resource, it is possible to prove falsifications after the fact. Western partners also understand this, so they are in no hurry to congratulate Gakharia on the victory, giving the opposition time to collect and bring facts to court.

Tamaz Khunjua, GSAC

illustrations