Was the Russian Election Fraudulent?

The Times today published a compelling report of first-hand observations of election fraud in Russia’s recent parliamentary elections. There are mounting protests; Secretary of State Hillary Clinton voiced “serious concerns” about the election and called for a “full investigation of electoral fraud and manipulation.”

But what if those first-hand observations were anomalous? What if the outcome for Vladimir Putin‘s United Russia Party, as disappointing as it was for him, truly represents the will of the Russian people?

A reader named Sergei Burkov, who identifies himself as “ex-head of Google Russia R&D Center” writes in with the following information:

Dear Sirs,

A great Freakonomics story tip:

Russian physicist Sergey Shpilkin managed to statistically analyze the results of the Russian Parliamentary elections, identify fraudulent component, and deduct it, to come up with the estimates of untainted results: Putin’s United Russia got 34%, not 51% as officially reported.

Done the same way as Google identifies (and deducts) fraudulent AdWords clicks and Palantir catches terrorists and money launderers.

I asked Burkov if any of Shpilkin’s findings were in English (the links above are in Russian) — and how the analysis had been done so quickly; the elections were held on Sunday. 

His reply:

No, I don’t think there is an English version yet. He had his software ready. He analyzed past results: esquire.ru/elections when the new results arrived, he just plugged the new numbers in. Raw data is published online in near real time on the website of the official Russian election commission.


These elections were going to be inevitably unfair for a number of reasons.

here are some examples:
1. Not all opposition parties were allowed to be registered and participate in the elections.
2. Only the government, and via that route Putin and his United Russia party, are allowed to broadcast on state-media (opposition parties are not allowed to spread their message).
3. websites of media, critical of the election process, were shut down or at least one radio station, two news websites, and a NGO.

The observers sent by the European Union also stated afterwards that (1) there are serious indications that fraudulent ballots were used, (2) the government "inappropriately intervened with the election process" and (3) the organization of the election was not done independently.

But for me the best part is the following article: http://www.nu.nl/politiek/2669299/rusland-hekelt-verkiezingswaarnemer-sp.html
unfortunately it is Dutch, so let me explain what it says:

Prior to the elections European observers were sent to Russia and they expressed two major concerns, which I have noted above, the unequal chances for parties to participate and manipulation of RESULTS (so AFTER voting). So basically they were saying that it would be unfair both prior and after the people had voted.

Now what did the Russians do when they were confronted with these allegations?

they sued the observers for "expressing critics on on-going elections" which isn't allowed BY LAW in Russia.

Again, you are Not Allowed, by law, to express critics on ongoing elections....



Of course it was fraudulent. How can anyone doubt it?!

Oleg Eterevsky

Sadly, even not taking into account the concerns mentioned by Nanno, these elections were far from being fair.

Another purely statistical analysis of the elections result is presented on the following graph: http://trv-science.ru/uploads/er.gif The horizontal axis represents the percentage of registered voters in a given district that participated in the elections, and the vertical axis is for the percentage of votes for a given party. The blue dots represent United Russia and the green ones — all the other parties combined. This is based on the officially published data. Basically, the most obvious explanation is that in the most fraudulent districts not only all the votes of the citizens that haven't come to the elections were counted towards UR, but also most of the votes for other parties were ignored and also added to UR.

I actually worked on these elections as an observer. My election precinct was actually relatively fair one. I suspect that the percentage of the fraudulent votes on it was no more than 10%. All of them were due to the people who were registered at several addresses and were payed for moving from one precinct to another and vote on each one. (This method is called "carousel" in Russia.) Unfortunately, I couldn't do much to prevent this because there were no means to prove that a given person already voted.



The results may be fraudulent, but most people (read Americans) reach that conclusion simply because the outcome is not to their liking. It is hard for them to fathom that someone like Putin can win with the margins reported in the official results.


Google Translator does a decent job of translating these pages to English. Here is a brief overview:

In most countries the distribution of the vote turnout by precinct more or less follows the bell curve (first chart at http://esquire.ru/elections show distribution for elections in Mexico, Poland, Bulgaria and Sweden). Russian election turnout distribution has a significant skew to the right, giving some support to the theory that some precincts have extremely high percentage of fraudulent ballots. Second chart at http://esquire.ru/elections shows that this skew has been growing in 2003-2008, presumably because the ruling party increasingly relies on this technique to keep itself in power.

It's not a great surprise that the precincts with high turnout are the same precincts where the distribution of the votes is heavily skewed towards the ruling party United Russia ("EP"). The first chart at http://podmoskovnik.livejournal.com/129843.html plots distribution of the votes EP received in Dec 4th elections (blue line) vs. the turnout. The higher the turnout, the greater the percentage of United Russia votes. The author attempts to break down EP votes into "normal" (light blue line) and "abnormal" (red line) components. Green line is the total votes for all other parties.

Another piece of evidence suggesting that something is not right is that the results are dramatically different in some neighboring precincts: http://www.echo.msk.ru/blog/varfolomeev/836376-echo/
This might be normal for US elections, where demographics can change dramatically when you cross a street, but it has never been the case in Moscow.


Doug Schaedler

The investigation of the Bowl Championship Series (BCS) could be very interesting in as much as the results seem strangely tied to influence rather then skillful results demonstrated on the playing field.

Russian Queen

I would not be surprised. I once taught a class to a bunch of Russian students. They hardly spoke English. Indeed, a few relied on translations by others in the class. Well, exam day and the senior member of the crew (a Russian md by profession) obviously used the situation to cheat and there was no way of my knowing/verifying one way or the other. I just refused to tolerate talking during an exam. He came close to failing and I never got called back to teach again. Apparently, the school was aghast. Cheating by an esteemed student was ok. So as I said, I would not be surprised. . I learned alot about our culture differences. The thing is- they just see things differently. Who knew?

Alex P

I wouldn't judge so bluntly about the whole population by one example. I'm a russian, and, although, I admit there are a lot of crocks and people who inclined to cheat, it doesn't mean that russians see things differently.
By the way, I'm studying right now in the UK with classmates from 38 different countries, and, despite cultural differences, people in out class treat cheating as horrible act. It's just humiliating to read that 'they see things differently'. No we don't. Some people do, but it doesn't automatically mean that everybody does so.

Regarding the elections, article publish in esquire, although based on preceding elections, is amazing and exactly elicit the reasons people took the streets after elections. Just try google translate to read http://esquire.ru/elections. It's worth it!

caleb b

Eh, I've seen worse.

The Sri Lankan incombent president won over his challanger by a reported margin that defies all logic, and his opponent (Fonseka) lost his own home town by 3 to 1. Then, Fonseka was thrown in prison right after the election and remains there today.


The funny thing about Russians is that they will always be Russians, corruption is in the eye of the beholder. While living in Egypt, in the 90s, the fine art of the payoff was alive and well. During a brief assignment to St Petersburg, the word corruption dosen't exist, the one with the most Rupels usually wins. The good side one will always know where one stands with Vlad.

Ilya Segal

Here is an upshot of the discussion in the Russian blogs. To infer the extent of the Russian election fraud from the publicly available data, it’s enough to apply high school level statistics and a bit of common sense. In “normal” elections, when we plot each party’s distribution of vote share in individual polling district, we should obtain a bell curve (i.e., Gaussian distribution). This is the normal result of adding up a large number of independent random effects. In contrast, look at the plot for the latest Russian election – the second graph on this page:
On the horizontal axis is the share of a party’s vote, on the vertical axis is the number of polling districts with this share (within a 0.5% bin). Different colors represent different parties. All the curves except the brown curve look almost like bell curves (with unusual peaks at the low end, to be discussed below). What strikes the eye is the strange behavior of the brown curve, which represents United Russia – the “party of the power”. While its upward-sloping part follows a bell curve, to the right of its peak it fails to follow the same curve, and instead extends all the way to 100%. Also, notice what Russian bloggers call “Churov’s comb” (Churov is the Head of the Election Committee): local peaks at “round numbers” starting from 50% all the way to 100%. Their explanation is that the heads of some local election districts rewrote their results to meet the round targets demanded “from the top.” (The number of such “idiots” is estimated to be about 1,000.) When making up the results, they gave low numbers to the opposition parties, which explains those parties’ unusual peaks close to 0. The smooth downward-sloping part of the brown curve is attributed to more sophisticated falsification, such as making up non-round numbers, ballot stuffing, and the use of hired “carousel voters” who voted dozens of times with the help of fake IDs.
These arguments are pretty convincing, as it seems pretty hard to come up with any alternative explanations for these curves. Together with the extensive reports of fraud from election observers, including videos, reports, and documents, the case is solid.

In conclusion: if we believe the case that United Russia’s actual vote share does follow a “normal” bell curve, and if we believe that the district results with low United Russia shares were not falsified, our estimate of their true vote share is shown with the blue curve on the second graph here:
Thus, if the election were not falsified, the vote for United Russia would have been about 32%, instead of the 50% officially reported.


Anton Nikolenko

I've written a short post based partially on the same source.



Here's an analysis done in English: http://samarcandanalytics.com/?page_id=39

Steve Coleman

An English version of Shpilkin's analysis can be found in the English-language version of Gazeta.ru, "Elections improbability" (12/23/2011). These same questions were raised after the previous parliamentary election in 2007. But a subsequent analysis: Stephen Coleman. 2010. "Russian Election Reform and the Effect of Social Conformity on Voting and the Party System: 2007 and 2008." Journal of the New Economic Society (Moscow), 5: 72-90.
(In Russian as "??????? ?????????? ????????????? ??????? ? ??????? ?????????? ???????????? ?? ??????????? ? ????????? ???????: 2007 ? 2008.") showed that the results can largely be explained by effects of social conformity, which has a stronger effect of Russian voters than in other countries. Although voting irregularities occurred, they did not have a substantial or material impact on the outcome. The article is available online at the website of the Russian New Economic Association, www.econorus.org or www.econorus.org/english.phtml.



I'd like to hear how "effects of social conformity" can explain spikes at 50%, 60%, etc. on 6th chart here: http://samarcandanalytics.com/?page_id=39