How do 137.84% of registered voters turn out to vote? They don’t.
In the aftermath of Kosovo’s “historic”, fleetingly violent elections on November 3, a dominant narrative seems to have emerged: The masked men breaking ballot boxes in Mitrovica were little more than a slight blemish on an otherwise near-impeccable event. Most reflections on election day now describe the attacks on polling stations in the north as a momentary “disruption” or “a minor disturbance”. The democratic process taking place elsewhere in Kosovo was, in the words of the Deputy Foreign Minister, “almost perfect”.
But were they? Not according to our analysis, which focused on Kosovo’s elections for municipal assemblies. (The elections also included a vote on local mayors). We found significant evidence of voter fraud, possible ballot stuffing, thousands of apparently missing votes, and an alarmingly high percentage of invalid ballots. After combing through some data, we contacted Kosovo’s Central Election Commission (CEC), but they refused our request for comment.
When we spoke with Florence Marchal, the spokesperson for the European Union Election Observation Mission – Kosovo 2013, she acknowledged that there were a “large number” of invalid votes, but declined to discuss any specifics before the Commission issues its full report in December.
Prior to presenting our findings, it’s important to reflect on Kosovo’s elections in December 2010, which were tarnished by accusations of “industrial-scale fraud”.
Various civil society organizations and political parties pointed out that over 90 percent of registered voters appeared to have cast a ballot at some polling stations. Given the fact that overall voter turnout was about 47.8 percent, the vast majority of election observers agreed that a 90 percent turnout was “highly unlikely”.
Re-elections were ordered for January in three municipalities, including two strongholds of Prime Minister Hashim Thachi, Skenderaj and Drenas, as well as Decan, and two polling stations in Malisheva and Lipjan. A dramatically lower voter turnout the second time around — by as much as 25 percent — seemed to confirm that the initial vote had been subject to significant manipulations.
However, not all the news was negative in 2010. The Kosovar Institute for Policy Research and Development (KIPRED), noted that “voter education has been improving during the years accounting for a continuously decreasing trend in the number of invalid votes.” According to the CEC, blank and invalid ballots accounted for a total of 9.05 percent of the votes in Kosovo’s 2007 elections, but accounted for just 4.6 percent of the vote in 2010.
However, in this month’s elections, over 11 percent of the total number of votes cast in the municipal assembly elections were invalid. (The CEC website currently lists the number of valid ballots as 659,116 and the number of invalid ballots as 81,782, for a total of 740,898 ballots.) According to the ACE Electoral Knowledge Network, three to four percent is considered the “generally accepted high level” for invalid ballots, while the global average proportion of invalid votes is “slightly less than three percent.”
Political analysts I spoke to in Prishtina said that the high percentage of invalid votes was likely the result of low voter education, lack of knowledge among the uneducated or elderly, and confusion among polling station commission (PSC) staff. This conclusion was similar to that of EU EOM, which glossed over the abnormally high percentage of invalid votes in its preliminary statement on the recent elections. Yet it’s certainly interesting that the opposite trend — a higher degree of voter education — was observed just a few years ago, and that the percentage of invalid votes in the most recent election was nearly four times higher than the international average.
What’s an invalid ballot? According to experts on electoral competition and invalid votes, a ballot can be considered invalid for different reasons, for instance: “The voter over-votes (i.e. casts more than one preference when only one preference is allowed) or takes an action that undermines the secrecy of the vote (e.g. she signs the ballot). The duty of an election officer is to invalidate any ballot on which the voter does not uniquely identify her preference.”
How do more than 100 percent of registered voters turn out to vote at a given polling station?
However, the high percentage of invalid ballots is far from the most troubling aspect of Kosovo’s allegedly “fair and free” elections. The irregularities we discovered were so extensive that we’ve decided to break this analysis up into a series. We’ll start with what seems to be the most obvious form of electoral fraud we found, and follow up with the less immediately identifiable manipulations in our second installment.
According to the CEC, a little over 47 percent of registered voters cast a ballot in the most recent elections. However, at a number of polling stations, a statistically improbable (or statistically impossible) percentage of registered voters turned out to vote. That is, more voters turned out to vote than there were registered voters for that polling station (that could mean dead people, double voting, etc.) Here’s a sample of some of the glaring irregularities of this kind found at various polling stations (identified by eight numbers and letters) across Kosovo:
Novobërdë, 1410D/01D: Allegedly had a voter turnout of 106.75%. According to the statistical information provided by the CEC, 154 of ballots were valid. 194 ballots were invalid.
Novobërdë, 1412D/01R: Allegedly had a voter turnout of 93.91%. According to statistical information provided by the CEC, 307 ballots were valid. 233 were invalid.
Novobërdë, 1416D/01D: Allegedly had a voter turnout of 121.14%. According to statistical information provided by the CEC, 59 ballots were valid. 90 were invalid.
Novobërdë, 1419D/01D: Allegedly had a voter turnout of 122.50%. According to statistical information provided by the CEC, 46 ballots were valid. 52 were invalid.
Junik, 3101C/04R: Allegedly had a voter turnout of 96.25%. According to statistical information provided by the CEC, 443 ballots were valid. 404 were invalid.
Ranilug, 3501D/03D: Allegedly had a voter turnout of 87.12%. According to statistical information provided by the CEC, 303 ballots were valid. 218 were invalid.
Obiliq, 1501E/06R: Allegedly had a voter turnout of 95.04%. According to the CEC, 507 ballots were valid. 393 were invalid.
Obiliq, 1507E/01D: Allegedly had a voter turnout of 96.95%. According to the CEC, 366 ballots were valid. 365 ballots were invalid.
Obiliq, 1512E/02D: Allegedly had a voter turnout of 87.95%. According to the CEC, 304 ballots were valid. 163 ballots were invalid.
Pejë, 1715C/02D: Allegedly had a voter turnout of 92.23%. According to the CEC, 178 ballots were valid. 202 were invalid.
Skenderaj, 2130B/02D: Allegedly had a voter turnout of 137.84%. According to the CEC, 450 ballots were valid. 457 valids were invalid.
Shtërpcë, 2303D/02R: Allegedly had a voter turnout of 95.96%. According to the CEC, 532 ballots were valid. 371 were invalid.
Ferizaj, 2534D/01R: Allegedly had a voter turnout of 99.60%. According to the CEC, 376 ballots were valid. 374 ballots were invalid.
Ferezaj, 2508D/02R: Allegedly had a voter turnout of 103.96%. According to the CEC, 393 ballots were valid. 473 ballots were invalid.
Kllokot, 3702D/01D: Allegedly had a voter turnout of 90.91%. According to the CEC, 201 ballots were valid. 149 ballots were invalid.
Gjakovë, 0229C/01R: Allegedly had a voter turnout of 98.47%. According to the CEC, 285 ballots were valid. 614 ballots were invalid.
Gjakovë, 0205C/10R: Allegedly had a voter turnout of 98.88%. According to the CEC, 345 ballots were valid. 451 ballots were invalid.
Gjilan, 0412D/01D: Allegedly had a voter turnout of 93.58%. According to the CEC, 120 ballots were valid. 521 ballots were invalid.
Klinë, 0809C/02D: Allegedly had a voter turnout of 93.16%. According to the CEC, 323 ballots were valid. 344 ballots were invalid.
Klinë, 0809C/01R: Allegedly had a voter turnout of 89.76%. According to the CEC, 361 ballots were valid. 489 ballots were invalid.
Podujevë, 1820E/03D: Allegedly had a voter turnout of 92.42%. According to the CEC, 220 ballots were valid. 341 were invalid.
Podujevë, 1806E/01D: Allegedly had a voter turnout of 93.60%. According to the CEC, 278 ballots were valid. 512 were invalid.
Prizren, 2067A/01R: Allegedly had a turnout of 92.44%. According to the CEC, 141 ballots were valid. 409 ballots were invalid.
Prizren, 2061A/01D: Allegedly had a turnout of 88.32%. According to the CEC, 330 ballots were valid. 313 ballots were invalid.
Suharekë, 2422A/03D: Allegedly had a turnout of 96.86%. According to the CEC, 178 ballots were valid. 316 ballots were invalid.
Malishevë, 3043A/01R: Allegedly had a turnout of 99.85%. According to the CEC, 320 ballots were valid. 330 ballots were invalid.
Malishevë, 3038A/02D: Allegedly had a turnout of 116.36%. According to the CEC, 469 ballots were valid. 43 ballots were invalid.
Partesh, 3601D/02D: Allegedly had a turnout of 93.84%. According to the CEC, 253 ballots were valid. 189 ballots were invalid.
We find these voter turnout statistics highly troubling and indicative of widespread electoral fraud (which we will elaborate on later). Earlier this week, Kosovo’s Deputy Minister of Foreign Affairs called us “devious liers [sic]” for questioning the rush to various conclusions about this month’s elections. We were even compared to Satan. However, we say the devil’s in the details — and in this instance, those details are in the individual polling station results. Stay tuned for more.
Photograph by Darmon Richter via flickr.
