Second Round of Elections
On 21 November, Georgia held the runoff in 17 single mandate constituencies. The opposition candidates withdrew from the contest and did not run electoral campaigns. However, their names remained on ballot papers. The turnout amounting to 26,29% (562,664 votes) was significantly lower than during the first round (56,11 %). Similar to previous malpractice, the runoff saw a number of unidentified persons mobilised near the polling stations across Georgia. Commonly identified as ‘coordinators’ they gathered outside the precinct stations or in vehicles with an apparent aim to record the number of voters. This largely contributed to creating a tense atmosphere and an air of surveillance over voters. In a number of instances, observers were prevented from exercising their duty, whilst at least four observers were threatened. There have been instances of multiple voting and of vote buying. Complaints addressing multiple voting have not been upheld.
Post-Election political standstill
The ruling party, Georgian Dream (GD), is set to assume 90 mandates in the 150-member parliament, taking 60 out of the total 120 seats in the proportional vote and existing 30 seats allocated through the majoritarian system. GD candidates were outright winners in the first round of October 31 parliamentary elections in 13 remaining single-mandate constituencies, while in the runoffs, in the absence of the existing candidates, they won all 17 races. To end the political standoff, the ruling and opposition parties held rounds of talks facilitated by the U.S. and EU Ambassadors, albeit with no success. The fourth meeting held on 9 December between the opposition parties and the ruling party saw differences over holding early elections and releasing ‘political prisoners’ whilst agreeing on most issues ‘in principle’. Just before the fifth round of discussions, the executive secretary of the GD suggested invoking mandates of the opposition party members based on their own initiative and pledged to strip off state funding and other privileges to the political parties who will boycott the parliament. He also unveiled a plan to set up a GD led enquiry commission with international observers and civil society to recount elections. In an unlikely case of more than one percent deviation in favour of the GD it was offered to call for early elections. The pledge did not go forward as revised PVT results narrowed the gap between the latter and the official result. It was also promised that regardless of the outcome of the negotiation, the major electoral reform will be delivered. The GD also suggested a draft law outlawing election registration of political parties whose leader does not possess a Georgian passport. The civil society saw it as an attempt to mislead the Georgian public and divert their attention from the main issues related to the 2020 Parliamentary Election. The measures were dubbed ‘punitive’, usually common for autocratic regimes. The first session of the 10th convocation of the parliament was convened on 11 December amid protests and in the absence of the opposition parties. On 15 of December, five opposition parties and blocks officially asked their party lists to be annulled. For the parliamentary session to be declared valid, it requires a quorum of 76 MPs which can be ensured even in the absence of the opposition. One party system, however, will significantly weaken the parliamentary oversight impeding passing certain laws in the absence of a necessary quorum.
Reasoning of decisions on electoral complaints and appeals
Throughout the election rounds, local citizen election observation missions largely voiced lack of trust in the election commissions and courts to handle election-related complaints impartially and effectively. The vast majority of complaints related to the recount of votes in relation to 109 precincts filed to the District Election Commission (DECs) were dismissed on merits, many without adequate investigation. Courts too mostly declined to order recounts. Out of 38 court cases brought before the courts, it granted recounts in 11 cases. Overall, this led to changes in five PECs and the annulment of results in one PEC altogether. Both election commissions and courts exceedingly relied on written statements by the registrars, largely circumventing standards for administrative proceedings. The Election Code of Georgia allows DECs to make changes in the PEC summary protocol of polling results. However, amending a summary protocol solely based on a statement of a representative of a lower-level administrative body fails to meet a high standard of administrative proceedings warranting a trust among election stakeholders.
The International Society for Fair Elections and Democracy (ISFED) voiced a lack of competence of judges in election related matters. As reported, judges fell short of adequately handling complaints or providing legal reasoning all together. The argumentations behind judgments were inconsistent, superficially touching on the subject matter whilst reasoning in some decisions have been identical. The courts appeared to have fully sided with the administrative bodies largely ignoring substantiated arguments provided by the election subjects. It was also assumed that courts failed to exercise its judicial oversight that would have led to a fairer outcome. Courts were usually late with submitting final decisions to the parties involved, substantially delaying appeal proceedings. Courts however have been particularly strict with admitting complaints with the slightest delay, striking them on inadmissibility grounds. Judicial review is critical in developing reasoned decisions and building stable administrative court practice for the resolution of electoral disputes.Reasoned decisions provide for the ‘verifiability of the decision’ and the ‘recourse to a remedy’ against the decision. The Venice commission emphasised the importance of the transparent nature of election dispute resolution systems, by ensuring procedures devoid of formalism. As suggested, only ‘transparency, impartiality and independence’ from politically motivated manipulation will ensure proper administration of the election process. It also follows that a very tight timeframe for complaint adjudication of two-day and one-day periods prescribed by the Election Code is largely prohibitive. It limited the number of complaints and appeals lodged by the citizen election observation missions. Time limits should allow the aggrieved parties to seek redress whilst the ‘Code of Good Practice’ recommends a time limit of three to five days as a reasonable timeframe. Tight timeframes also give raise to issues related to the burden of proof. Administrative courts enjoy an inquisition role in the dispute resolution allowing them to obtain evidence by its own initiative.Within tight timeframes, however, courts lack an opportunity to obtain, examine and request evidence on its own. The Venice commission finds the burden of proof in electoral disputes to be important, albeit problematic due to the extremely limited time for adjudicating the complaint or the appeal.
‘Rigged’ Election Claims
All opposition parties that passed the one percent threshold in the first round refused to enter the new parliament, boycotting the outcome of the ‘rigged’ October 31 Parliamentary Elections. A statistical analysis of the official results conducted by the EPDE expert Tamaz Khunjua confirms some of the concerns of the observers and indicates a possible impact of the irregularities on the overall election result.
Differences between the majoritarian and proportional vote
The statistical analysis of the results from all precincts in Georgia during the first round showed in several cases significant differences between the majoritarian and proportional vote. The methodological assumption in this analysis is that the divergence between votes cast for the majoritarian candidates of a particular party and votes for the lists of the same party should not be significant. The analysis shows that the divergence in almost all 3,847 polling stations is indeed very small. However, there are several exceptional polling stations located in three districts where the ruling party's majoritarian candidates received a minimum advantage to pass in the first round. In those polling stations the difference between the results of the ruling party in proportional elections and their candidates in majoritarian elections were proportionally high. Among those districts were:
An analysis of the results in some polling stations in those three districts shows the source of these divergences:
District #13 (Gardabani): In the polling station number 13.22.69 (Marneuli Kindergarten No 4) the ruling party received only 216 votes while its majoritarian candidate received more than double as much - 450 votes (a difference of 234 votes). It is important to note that in the initial voting protocol there were only 252 votes reported for the majoritarian candidate, however, an addendum was made to the protocol on November 1 in which this number was corrected to 450. Similar differences were identified in this district e.g. in the polling stations no. 13.22.03 (Marneuli public school no 8: where GD received 249 votes while its majoritarian candidate gained 330 votes - difference of 81 votes), no. 13.22.34 (Marneuli, village Kapachkhani: 254 votes for GD while 327 for its majoritarian candidate - difference of 73 votes) and in no. 13.22.37 (village Keshalo: 246 votes for GD while 365 votes for its majoritarian candidate - margin of 119 votes).
District #25 (Abasha): In polling station number 25.65.02 (Martvili) the ruling party received 398 votes in the proportional elections while its majoritarian candidate received 506 votes. The margin amounted to 108 votes. A similar situation was identified e.g. in the polling station number 25.65.01 (Martivili) where the governing party received 310 votes, whereas its majoritarian candidate received 431 votes - difference of 121 votes.
District #29 (Kobuleti): In polling station number 29.81.55 (village Kvirite), the government party received 226 votes in the proportional election while its majoritarian candidate received 337 votes in this polling station - difference of 111 votes.
These diverging results do not necessarily indicate an election fraud, however, the conduct of elections in the mentioned districts could be a subject of further investigation by international and domestic election experts.
Correlation between the outcome for the ruling party and an increased turnout (Kiesling-Shplikin method)
The outcome of the proportional elections was analysed using the Kiesling- Shplikin method which shows the correlation between turnout and the results of the parties. When using this method to assess election results in well-established democracies, the diagram would show a so-called “normal distribution” which means that the majority of polling stations would indicate average turnout by average results of the contenders. While unproportionally high or low turnout by unproportionally low or high results of the contenders would indicate irregularities and a need for further scrutiny. According to numerous studies on elections in different countries based on this method, the opposition usually gains more votes in polling stations with high voter turnout. In the case of the 31 October election in Georgia, however, the trend appears to be the opposite, especially in the eastern (election districts no. 9 - 16) and western (election districts no. 17 - 30) parts of Georgia. The diagram below reflects the results of the ruling Georgian Dream party (blue points) and all opposition parties together (red points) versus turnout.
Each point corresponds to the existing precinct election commission. The analysis of the results from those regions shows the following trend: the higher the turnout the better scored the ruling party (positive slope towards the upper right corner) and the worse the results of the opposition (negative slope towards the lower right corner). While this trend requires further scrutiny, it is alarming that irregularities have been reported by observers from polling stations with particularly high turnout.
As an example, the algorithm identified the precinct election commission no. 28.79.19 in Batumi being marked with a high number of reported irregularities. The number of votes cast in this polling station exceeded the number of registered votes. In this case, the results were successfully appealed at the District Election Commission (DEC).
The analysis based on Kiesling-Shpilkin methodology would consider results only from polling stations with a “normal distribution” of votes (the majority of polling stations) while rejecting figures recorded in polling stations with significant deviations from the average. If we assume that the deviations from the normal distribution identified during the statistical analysis of the results from all polling stations in Georgia are a consequence of an external interference in the electoral processes, then the estimated intervention would amount to approx. two to three percent in favour of the ruling party.
Recommendations to the Georgian authorities
Author: Mariam Uberi
Statistical analysis: Tamaz Khunjua