ISFED Board believes that the fault discovered in the parallel vote tabulation is a human error but the Executive Director is responsible for the delayed communication of the error with the public
On December 25, at 13:00, briefing of ISFED Board was held at the office of ISFED, where board members Baia Pataraia and Tamara Sartania presented internal inquiry report and recommendations of the board with regard to future steps.
International Society for Fair Elections and Democracy (ISFED) conducted a comprehensive monitoring of the entire process of the October 31, 2020 parliamentary elections, including through long-term and short-term observers. Monitoring tools of the organization include parallel vote tabulation (PVT), which ISFED has been implementing for all national elections since 2003.
PVT methodology entails stationing short-term observers at representative polling stations randomly selected nationwide and processing information reported by them. PVT methodology ensures monitoring of the entire Election Day process and identification of irregularities, as well as quick transfer of voting results to the PVT data center and their processing.
On November 1, 2020, ISFED published preliminary PVT results, which differed from the Central Election Commission returns for three parties (the PVT results differed from the CEC returns for other parties as well but within the margin of error).
Audit performed by the organization in November discovered an error in the vote counting formula. In particular, invalid votes were mistakenly included in the counting of votes. After correcting the error, party vote shares were changed only fractionally, moving closer to the CEC returns.
ISFED published a statement about revised results on December 11, giving rise to additional questions in the society. Questions were raised about impartiality and professionalism of the organization and suspicions were expressed about pressure against ISFED and its executive director. Long-standing reputation of the organization was called into question.
In view of high public interest, upon request of ISFED board member, at the extraordinary meeting on December 12, the board voted to approve a commission for investigating causes of the error identified in PVT.
The purpose of the commission was to study possible causes of the error identified in the parallel vote tabulation during the 2020 parliamentary elections and examine any measures taken by the organization management to address the problem.
The commission conducted the inquiry on December 12-20. It interviewed ISFED Executive Director, Deputy Director, employees involved in PVT process, ISFED upper and middle management, as well as one of the experts of the National Democratic Institute (NDI), who was personally involved in the 2020 PVT process. The commission also studied relevant documentation.
ISFED has conducted parallel vote tabulation for every national election since 2003. In addition, PVT methodology is used in other countries as well and improved every year, among them, based on experience of ISFED and other countries.
NDI provides technical assistance to PVT conducted by ISFED. For the 2020 elections, the assistance entailed support of PVT implementation and analysis, including preparing SMS and incident base, as well as managing PVT center and database on Election Day, processing and analyzing the received data with ISFED.
The commission has found that SMS and incident base is updated by NDI experts, based on information discussed with ISFED. This information is used, for example, to revise vote tabulation formula in certain countries, as well as to prepare logical problems for identifying irregularities in the election process (e.g.,imbalance in protocols, inaccuracy in voter turnout, etc.).
The commission has found that ISFED provided to NDI experts a written list of logical tasks, however the list did not include Georgia-specific formula for counting votes obtained by election subjects. The commission could not establish without a doubt whether or not the issue of counting invalid ballots was discussed verbally during meetings.
Before Election Day, in order to test SMS and incident base of PVT, ISFED conducts Election Day simulation, however the process does not entail simulation of counting of votes received by election subjects. Therefore, according to information provided to the commission, ISFED could not have discovered the error in vote tabulation formula even during testing.
The commission also considered the possibility of any unscrupulous manipulation with the software; however, it did not find a single corroborating fact. The commission therefore believes that assumptions about deliberate discrepancy between PVT data and CEC returns are ill-founded.
Having conducted the interviews and studied relevant documentation, the commission finds that the error discovered in the vote tabulation formula is a human error, and ISFED PVT Coordinator/Deputy Director and NDI Washington Office experts were equally responsible for preventing it.
Additionally, the commission has found that initially, only the executive director and the deputy executive director knew about the error discovered in the formula in ISFED. According to them, the management recognized the necessity to make the information public, however because the maximum error was 1.7%, which according to them did not influence distribution of mandates – publishing the statement was postponed until the audit process was completed, when all other possible data would also have been verified. Later the executive director made a decision to publish final PVT results with the interim report of the parliamentary elections.
According to the executive director, on December 9 she learned that in the process of negotiations with opposition, the ruling party was manipulating with PVT results. In addition, on December 11, media reported that ISFED executive director was pressured. In response, the executive director made a decision to publish updated PVT results.
The commission believes that the executive director failed to consider the existing political context and she did not communicate to public information about the error found in the vote tabulation formula, jeopardizing reputation of the organization and credibility of PVT methodology in general. In addition, the commission did not find any attempts of pressure or intimidation against the executive director, which would have forced her to hide this information.