"Smart Voting" as One of the Determining Factors in the 2019 Moscow City Duma Election Results

Lyubarev A.E.

Abstract

The paper discusses factors that influence elections and their role in the 2019 Moscow City Duma election. The author conducts a critical analysis of models used by other researchers in assessing the "Smart Voting" effect. The analysis is followed by the author presenting his own model and subsequently inferring that "Smart Voting" produced a 14% effect in constituencies with an available tactical voting choice between CPRF and A Just Russia. The paper argues that "Smart Voting" played a catalytic role in the opposition's success in eight single-member constituencies and presumably three more. Two constituencies indicated cases of incorrect detection of the supported candidate, which resulted in the opposition's defeat. In conclusion, further use of "Smart Voting" technology in the election is discussed.

---

We are continuing the discussion of the impact of "smart voting" the idea and "Smart Voting" the campaign on the vote returns of the 2019 Moscow City Duma election, which began in papers by Andrei Buzin [3] and Ivan Bolshakov with Vladimir Perevalov [1] published in this issue. We are also considering earlier publications of Boris Ovchinnikov [12] and Ivan Bolshakov with Vladimir Perevalov [2] on the same topic. To differentiate between the idea and the campaign, we shall use lowercase and uppercase letters respectively.

Factors influencing vote returns

When assessing the impact of any factor on vote returns, one has to realize that a great number of factors operate in combination and they all have to be considered.

Below are the factors that affect vote returns of an individual candidate according to our data.

1. Candidate's activity during the campaign.

2. Candidate's characteristics that a voter can see on the ballot: sex, age, employment/occupation and (in certain cases) ethnicity, which can de deduced from the candidate's last, first and patronymic names. This is especially important for voters who did not see the campaign.

3. Party affiliation.

4. Electoral specifics of a constituency, i.e. the balance of political powers.

5. Having other candidates with similar characteristics (both political and personal).

6. Support from public opinion leaders.

Let us start by checking if these factors applied during the 2019 Moscow City Duma election campaign.

1. Assessing the level of activity of various candidates during the campaign is quite difficult. In order to paint an objective picture, we need to monitor the campaign more thoroughly in every constituency with the help of a thought-out unified method. It is unlikely that anyone did this kind of monitoring. We only have descriptions acquired through limited monitoring activities at our disposal [7].

We can try to use data on total expenditures of candidates on the election campaign as an objective measure. However, one has to exercise great care when considering this data. There is a possibility of the numbers being either understated (prompting illicit spending) or overstated (prompting embezzlement), as well as simple ineffective expenditure. Besides, a large number of effective campaigning methods (meeting voters in small groups, candidate's own participation, volunteer help, canvassing) requires little spending, which is probably the most important factor. The analysis did not reveal any consequential correlation between expenditures and results for various more or less homogeneous groups of candidates.

In any case, let us try to compare the data we have. The data on campaign expenditures and vote returns for candidates discussed below are shown in Table 1. As for campaigning activity, most administration independents, CPRF candidates and liberal candidates indicated relatively high levels according to available reference material [7].

Among the least active administration independents challenging for the win were Sergei Zverev, Andrei Medvedev, Irina Nazarova, Yevgenii Nifantyev and Anastasia Tatulova. That said, Tatulova and Zverev showed the lowest results in the group. Nifantyev's result is in the lower half as well, although Medvedev and Nazarova showed relatively high results.

Expenditures of all administrative independents challenging for the win were quite high. Sabina Tsvetkova, Nikolai Tabashnikov, Roman Babayan, Aleksei Shaposhnikov and Anastasia Tatulova indicated the lowest expenditures. With that in mind, Tabashnikov's result is very low while Babayan's result is the highest among independents by contrast. Tsvetkova got a relatively low result, whereas Shaposhnikov's was relatively high.

We can single out two administrative independents who served a technical purpose. Yekaterina Kopeikina's campaigning activity was barely noticeable. Vera Shevchenko's – rather weak. However, Kopeikina only spent 183 000 RUB (~2 620 USD) (with about 76 000 RUB (~1 100 USD) spent on collecting signatures) while Shevchenko's expenditures amounted to 28 000 000 RUB (~400 000 USD). Kopeikina's result is 6.8% and Shevchenko's is 17.9%.

Among the least active CPRF candidates were Dmitry Agranovsky, Aleksandr Vidmanov, Yulia Gladkova, Yelena Gulicheva, Yurii Dashkov, Sergei Desyatkin, Lyudmila Yeryomina, Andrei Ispolatov, Konstantin Lazarev, Aleksandr Potapov and Nikolai Stepanov. As for the expenditure, most candidates had similar numbers ranging between 800 000 and 1 400 000 RUB (~11 500 and 20 000 USD). Exceptions included Olga Frolova, Sergei Desyatkin, Dmitry Agranovsky, Anastasia Udaltsova (expenditure less than 800 000 RUB (~11 500 USD)); Lyudmila Nikitina, Yelena Yanchuk, Nikolai Gubenko, Vadim Kumin and Aleksei Melnikov (expenditure over 1 400 000 RUB (~20 000 USD)).

Among the CPRF candidates, Agranovsky, Ispolatov and Yeremina showed the lowest results. Fourth to last, however, was Aleksei Melnikov, who is the record holder for expenditure. Neither of these candidates was endorsed by "Smart Voting" strategy. The next five spots before Melnikov are occupied by four candidates not endorsed by Alexey Navalny (including Konstantin Lazarev and Aleksandr Potapov) as well as Vladislav Zhukovsky, who ran against Roman Yuneman in constituency no. 30. In this case, we still have to figure out how low activity impacted the low result and what was the cost of not being endorsed by "Smart Voting" strategy. However, the lowest results among those endorsed by Navalny were produced by Vidmanov and Dashkov (that is besides Zhukovsky). This allows us to discuss the correlation between activity and result.

At the same time, the not-so-active Frolova, Gladkova and Desyatkin got moderate results, whereas Gulicheva and Stepanov's results are relatively high. On the one hand, we may observe the high results of Nikitina and Yanchuk, both of whom were financially secure and campaigning actively.

The most active A Just Russia candidates were Yevgenii Borovik, Vladimir Zalishchak, Boris Kagarlitsky, Mikhail Timonov and Georgii Fedorov. As for expenditure, several candidates indicated zero (including Magomet Yandiyev, constituency no. 45 winner), whereas several more spent less than 100 000 RUB (~1 500 USD) on their campaigns (including Aleksandr Solovyov, constituency no. 3 winner). At the same time, the expenditures of 11 candidates went 1 000 000 RUB (~14 000 USD). These candidates include Mikhail Timonov, Georgii Fedorov, Valery Danilovtsev, Sergei Vasilyev, Andrei Medvedkov, Natalya Yankova, Vladimir Kalinin, Alexey Glek, Boris Kagarlitsky, Aleksandr Luchin and Aleksandr Romanovich.

Besides Solovyov and Yandiyev, Timonov and five other candidates endorsed by "Smart Voting" showed the best results among A Just Russia candidates, including the financially secure Vasilyev, Glek and Kalinin as well as Ilya Lifantsev and Arkadii Pavlinov, both of whom did not campaign much.

The best results among A Just Russia candidates not endorsed by "Smart Voting" went to Romanovich, Panina, Borovik and Fedorov. It may seem that in this case, the impact of a candidate's campaigning activity is showing. However, another factor was in effect: none of the four candidates ran against either an administrative independent (Panina, Romanovich, Borovik) or a CPRF candidate (Fedorov). The results of the actively campaigning Kagarlitsky and Zalishchak were more modest.

The relatively high results among other groups of candidates can be observed for actively campaigning independents Mikhail Konev, Sergei Malakhov, Marya Marusenko, Mikhail Menshikov, Aleksei Sobolev, Anton Tarasov, Aleksei Shkolnikov as well as Daria Mitina, A Communists of Russia candidate.

A review of this analysis indicates that a candidate's campaigning activity had seemingly affected the result in some way, yet we do not have any evidence it had been a deciding factor in any of the cases.

2. The role of a candidate's characteristics is best tested on the example of the 40 administration nominees running for the Duma seats. It turns out that the candidate's sex did not make any difference: women's average result amounted to 36.0% while men's to 35.2%.

The occupation seemingly did, however, although it is difficult to make any reliable conclusion because of the small number of candidates in each group. Re-elected Moscow City Duma deputies showed the best average result (16 candidates, 37.5%), as did journalists (5 candidates, 37.3%). The average for directors of state-funded institutions and social activists is a bit lower (8 candidates, 35.8%). Business executives follow with an even lower average (4 candidates, 33.9%). Heads of higher education institutions displayed the lowest average, however (4 candidates, 27.9%). At the same time, the highest result among rectors and vice-rectors is 32.4% while 31.9% is the lowest among state-paid workers and social activists.

It is also likely that the lowest result (20.8%) among administration candidates belongs to Kirill Nikitin, whose occupation on the ballot did not look all that becoming: he was billed as the director of the division of tax services for state government bodies, city government bodies and public sector authorities of the branch office of PricewaterhouseCoopers Russia B.V.

3. The role of party affiliation is quite easy to demonstrate. Communists of Russia candidates (except one – in the constituency without a CPRF candidate) gained between 2.7% and 11.5%, LDPR candidates (except two – in the constituencies without administration independents) gained between 4.1% and 12.8% with CPRF gaining the lowest at 11.8%.

We cannot really talk about party affiliation of administration candidates, although mayor's office endorsement identified them as "party of power" representatives to a certain extent. In any case, this factor may be referred to the "Support from public opinion leaders" category (item 6 in our classification).

We do have an isolated example, however. Constituency no. 21, which the mayor's office "ceded" to CPRF representative Leonid Zyuganov, had Vera Shevchenko, an executive secretary of the local United Russia branch. She did not make any claims to victory and did little campaigning. Nevertheless, the ballot affiliated her with United Russia (as place of employment), so she gained 17.9% because of that.

4. Analysis of electoral specifics of constituencies yielded interesting results. Andrei Buzin [3], along with Ivan Bolshakov and Vladimir Perevalov [2] attempted to address these specifics through comparison with the 2014 Moscow City Duma election results. We believe this is not an acceptable way to address the issue, which we intend to clarify further.

Analyses of the specifics of electoral geography were previously carried out by Vladimir Kolosov and Olga Vendina [4], followed by the authors own analysis based on the elections of 1995–2000 [10: 338–343]. Moscow's electoral geography changed significantly between 2004 and 2011 due to massive amounts of electoral fraud, but came back to normal in 2012 [9]. It may be briefly described as follows: districts outside the Moscow Automobile Ring Road (MKAD) and the adjacent districts mostly tend to support the "party of power"; central districts as well as those positioned along Kutuzovsky, Leninsky, Leningradsky, Vernadsky and Mira Avenues tend to support both left-wing and right-wing opposition; the so-called "Red Belt" (meaning the districts with relatively strong left-wing opposition) includes districts close to Moscow Central Circle (MCC, part of Moscow railway); on average, South-Western, Western and North-Western Administrative Okrugs are opposition-leaning as well, Southern and South-Eastern are more pro-government, whereas the districts of Eastern and North-Eastern Okrugs are close to moderate.

However, the delimitation of constituencies in the 2014 and 2019 Moscow City Duma election was quite "sneaky". Many districts of the "old" Moscow (22) and both administrative okrugs (AO) of the New Moscow were divided between single-member constituencies. Besides, one constituency often incorporated pro-government and opposition districts. As a result, we needed to do additional calculations to address the electoral specifics of single-member constituencies.

We believe that the 2016 legislative election results by party lists are the best reference in assessing electoral specifics of Moscow's territories. All major parties ran in this election (as opposed to the 2018 presidential election), party representation was the same all over Moscow and voting was strictly party-based (having a candidate in regional groups or single-member constituencies was a minor concern), which is why this election may serve as an indicator for the level of support for any given party in different parts of the metropolis. Turnout differences between the 2016 and 2019 election may be discarded in this context.

Since we had the data on vote returns across all districts and polling stations, we were able to work out a table on returns (Table 2).

As we can see, the differences across constituencies are obvious and mostly fall within the previous conclusions. United Russia's highest result (49.9%) was in constituency no. 30 (South Chertanovo District, which is adjacent to MKAD, as well as the neighboring Central Chertanovo, both part of Southern AO) while the lowest (29.7%) was in constituency no. 37 (districts of South-Western and Western AOs near Vernadsky and Leninsky Avenues). CPRF's results are not as skewed, with the highest standing at 15.9% in constituency no. 21 (parts of Vykhino-Zhulebino and Ryazansky Districts, which belong to South-Eastern AO; a result of Andrei Klychkov being elected there in 2014) and the lowest at 11.3% in constituency no. 30. The Russian United Democratic Party (RUDP) Yabloko got its highest result (17.2%) in constituency no. 37, and the lowest (5.2%) in constituency no. 29 (Western and Eastern Biryulyovo Districts, both adjacent to MKAD, as well as part of Tsaritsyno District; all districts belong to Southern AO). A Just Russia received its highest result (9.6%) in constituency no. 7 (the districts of Northern AO adjacent to MKAD) and lowest in constituency no. 38 (New Moscow with a fraction of Western AO).

The next intriguing thing to figure out is whether a connection exists between the 2016 and 2019 vote returns. For this purpose, we used correlation analysis. However, it was obvious there were several special points that created distortions. These are constituencies no. 8, 20, 21, 37 and 43 – the ones either without administration independents or with them serving a technical purpose – as well as no. 3, 14 and 45 without CPRF candidates. We have therefore calculated the correlation between vote returns in 2016 and 2019 (as a percentage of the number of voters) for both 45 and 37 constituencies, the latter excluding the constituencies listed above. The results are displayed in Table 3.

As the table indicates, the correlation between the 2016 and 2019 results is insignificant for nearly all parties except LDPR if special points are discarded. However, compared to 2016, the changed vote returns of 2019 already indicate a quite a significant correlation with both 2016 results of RUDP Yabloko and CPRF (positive), and 2016 result of United Russia (negative). The correlation between the 2016 result of RUDP Yabloko and the 2019 results of "Smart Voting" candidates is also traceable. This means that the government and the opposition had respectively lost and gained more votes in the more opposition-leaning constituencies than in the initially less opposition leaning. In other words, political field became more polarized.

5. The impact of having or lacking candidates with similar characteristics or shared electorate can be illustrated through some notable examples. For one, in two constituencies without administration independents, LDPR candidates gained 19.1% and 20.1%, whereas in other constituencies their result did not go over 12.8%. In two constituencies without CPRF candidates, A Just Russia candidates gained 35.0% and 38.1% of votes despite the lack of active campaigning on their part – a similar result for JR (36.4%) can only be found in constituency no. 16, where Mikhail Timonov ran an active campaign; their results were more modest – between 25.0% and 28.4% – in the remaining five constituencies with JR candidates being endorsed by "Smart Voting". In the constituency without a CPRF candidate, Communists of Russia gained 20.6% of votes, whereas their results did not go over 11.5% in the remaining constituencies.

Boris Ovchinnikov [11] clearly demonstrated the impact of having candidates from Communists of Russia on the results of CPRF candidates. His calculations indicate that CPRF candidates endorsed by "Smart Voting" received an average of 36.2% in the 23 constituencies with CPCR candidates in the ballots (exclusive of the special constituency no. 30, where the result was notably lower). The median result in the same constituencies amounted to 35.2%. The average result amounted to 41.3% across the seven constituencies with CPRF candidates endorsed by "Smart Voting" but without the CPCR spoiler, while the median amounted to 42.2%. These calculations helped draw a conclusion that on average, CPCR candidates took away about 5% from CPRF candidates.

Assessing the impact of candidates with similar personal characteristics (last name, sex, and occupation) is more challenging. For example, CPRF's Pavel Tarasov was running against an independent Anton Tarasov in constituency no. 24. The latter gained 7.4% of votes, which is a decent number for a technical candidate. Nevertheless, Pavel Tarasov's result was quite high (39.0%), but Aleksei Balabutkin on the other hand, a Communists of Russia candidate in this constituency, achieved the lowest result among his fellow party members (2.7%). It thus seems like that Anton Tarasov took away more votes from Aleksei Balabutkin than he did from Pavel Tarasov.

6. The most striking example of a public opinion leader influencing the election is Alexey Navalny's "Smart Voting". Experts do not dismiss the fact of its impact in itself: the debate surrounds the impact scale. We will discuss this issue in further sections of this paper.

As we have observed, many factors affect vote returns. For this reason, if possible, it is essential to eliminate all other factors when trying to assess the impact of an individual factor.

The challenges of assessing the impact of "Smart Voting"

First, while considering the impact of various factors, one has to infer that giving assessment based on the averaged data from all 45 constituencies (or from most of them) is unreasonable. This essentially involves the differences in candidate rosters across constituencies, which, as we can see, translates into vote returns.

Only 14 constituencies share a similar selection of candidates: an administration independent, CPRF, LDPR, A Just Russia, Communists of Russia. These are constituencies no. 1, 4, 7, 9, 10, 12, 13, 22, 26, 27, 31, 34, 38, and 39. We assume that it is possible to assess the impact of the other factors within this group precisely.

Other constituencies have to be divided into smaller groups. That said, five constituencies had only an administration independent, CPRF, LDPR and A Just Russia on the ballots (no. 5, 11, 17, 28, 29). Ballots of seven constituencies included an administration independent, CPRF, LDPR, A Just Russia and Communists of Russia, as well as either technical independent or a candidate from a less popular party (no. 16, 24, 32, 33, 35, 36, 40). Two constituencies included an administration independent, CPRF, LDPR, A Just Russia and a technical independent (no. 2 and 18). Another two constituencies included an administration independent, CPRF, LDPR and Communists of Russia (no. 19 and 23), while in two more these four nominees were joined by a technical independent (no. 15 and 41).

The rosters of 13 constituencies are unique; these constituencies are no. 3, 6, 8, 14, 20, 21, 25, 30, 37, 42, 43, 44 and 45.

Second, we believe that comparing the 2014 and 2019 vote returns is inefficient. This is also due to the differences in candidate selections already described by fellow researchers [3; 1]. Suffice it to say, RUDP Yabloko candidates were on the ballots of 44 constituencies in 2014 and only on three in 2019. On the other hand, Communists of Russia candidates were not registered in any constituency in 2014, whereas in 2019 they were represented in most (31).

Still, even the parties that were represented in most constituencies during both campaigns ran with extremely varied selections. In the 2014 and 2019 election, an overall of 27 candidates ran in the same constituencies. Of them, 13 were administration candidates (Nadezhda Perfilova, Larisa Kartavtseva, Aleksei Shaposhnikov, Andrei Metelsky, Anton Molev, Inna Svyatenko, Lyudmila Stebenkova, Kirill Shchitov, Stepan Orlov, Sergei Zverev, Lyudmila Guseva, Aleksandr Semennikov, Olga Sharapova), 6 CPRF candidates (Pyotr Zvyagintsev, Nikolai Zubrilin, Yelena Gulicheva, Pavel Tarasov, Nikolai Gubenko, Yelena Shuvalova), two LDPR candidates (Maya Galenkina, Anton Yurikov), two A Just Russia candidates (Mikhail Timonov, Sergei Vasilyev), one RUDP Yabloko candidate (Maksim Kruglov), an independent candidate Mikhail Menshikov as well as Ilya Sviridov, who ran as an A Just Russia candidate in 2014, but went independent in 2019. The last candidate in this group is Viktor Gogolev, who ran independent in 2014, but as a Communists of Russia candidate in 2019.

As a result, candidate personnels were completely different in 23 constituencies (i.e. in more than half). Seventeen constituencies shared a common candidate and only five shared two.

There is no doubting that candidate personnel plays an important role. For instance, Andrei Buzin [3] indicates the extent to which Yevgenii Bunimovich and Segrei Mitrokhin improved the results of Sergei Grigorov and Sergei Ivanenko respectively. Still, comparing Yevgenii Bunimovich, who is quite well-known and filled an important position, to a much less known Sergei Grigorov is not entirely acceptable. Even Sergei Ivanenko, despite his political experience, turned out to be a weak candidate by receiving less than the average result for RUDP Yabloko in a clearly opposition-leaning constituency. That said, comparing his result to that of an actively campaigning Sergei Mitrokhin does not seem acceptable either.

The third aspect concerns the specifics of the "Smart Voting" campaign. As the author pointed out in his preface to the discussion [8], "smart voting" is essentially a form of "tactical voting", and voters often use it intuitively. At the same time, after being pushed forward by Alexey Navalny, the idea of "smart voting" could incite the opposition-leaning voters to such electoral behavior irrespective of recommendations given by "Smart Voting" the campaign.

Note that opposition-leaning voters had limited choice: candidates from CPRF, A Just Russia and RUDP Yabloko as well as opposition independents. However, RUDP Yabloko candidates and/or opposition independents could only be found in five constituencies. Liberals would have supported Yevgenii Bunimovich, Daria Besedina and Sergei Mitrokhin even without Alexey Navalny's assistance. In the meantime, Alexey Navalny's campaign mostly influenced the liberal electorate. The question of whether this campaign could urge the left-wing electorate to vote for Yabloko remains open. The situation in constituencies no. 14 and 30 requires a discrete analysis.

Opposition-leaning voters did not have any choice in constituencies no. 3 and 45, where neither CPRF nor liberal candidates stood for the office. A similar scenario played out in constituencies no. 15, 19, 23 and 41, with neither A Just Russia nor liberal candidates running for office. An even trickier situation developed in constituency no. 44, where A Just Russia's candidate Ilya Sviridov ran with some endorsement from the mayor's office while apparently managing to have kept some support from the opposition.

As a result, the situation in the 33 constituencies, where the opposition-leaning voters could choose between CPRF and A Just Russia seems more noteworthy. "Smart Voting" campaign endorsed a CPRF candidate in 27 out of 33 constituencies, and an A Just Russia candidate in the remaining six (no. 13, 16, 26, 28, 35 and 38).

The 2014 and 2019 results-based models

The model developed by Ivan Bolshakov and Vladimir Perevalov [1; 2] is based on an attempt to separate the effects of "smart" (tactical) and protest voting. According to their concept, tactical voting is premised on rational choice (voting for the candidate with best chances of winning to prevent the election of the administration candidate) while protest voting is rooted in a social and psychological foundation and dictated by a desire to punish the ruling party and demand change.

Instrumentally, protest voting was assessed based on the results of a candidate decreasing in the 2019 election compared to the 2014 election. Tactical voting was assessed based on votes overflowing to an alternative candidate (i.e. endorsed by "Smart Voting") from other non-administration candidates in the 2019 election compared to the 2014 election (where a more successful non-administration candidate was referred to as alternative) [2]. This approach has a number of significant shortcomings. As was pointed out before, comparing the 2019 vote returns to those in the 2014 election is difficult because of an entirely different candidate selection. The approach of Ivan Bolshakov and Vladimir Perevalov looks especially flawed wherever the order of candidates had changed compared to the previous election. That said, they refer to Mikhail Timonov as a "minor" candidate in 2014, whereas in 2019 he is referred to as alternative. As a result, when calculating the tactical voting effect for constituency no. 16, they factor in the "vote overflow" from Timonov to Timonov, but fail to do the same with votes overflowing to Timonov from the communist candidate Dmitry Parfyonov. Bolshakov and Perevalov therefore find that tactical voting effect amounts to only 2.1%. In addition to the above, the model fails to answer the following questions: 1) why the 15% of votes lost by administration candidate Anton Molev went to Timonov; 2) why the correlation of votes for CPRF and A Just Russia candidates was at 1.38 in 2014, but at 0.48 in 2019.

As we also established before, trying to make calculations for all 45 constituencies (or even 43 of them) using a unified and highly formalized model is unreasonable. The results produced by Ivan Bolshakov and Vladimir Perevalov [2] clearly attest to that. Their protest voting effect ranges between 0 (in 8 constituencies) and 26.3% (in constituency no. 17). However, if we think of protest voting as an actual type of electoral behavior and not just a formal indicator, it cannot range as much in the context of a considerably homogeneous metropolis. It is obvious that dismissing the factor of administration candidates' personal characteristics while calculating the level of protest voting is unacceptable.

It is possible to consider the main idea of the model developed by Ivan Bolshakov and Vladimir Perevalov controversial as well, meaning the idea that it is possible to separate tactical voting from protest voting. The very characterization of protest voting, which is based on foreign studies, is not exactly applicable to Russian context, where the majority of voters do not have a steady party preference and the choice is often determined by personal rather than party preference, and a non-partisan candidate can get a high number of votes. The assertion that protest voting does not pursue the goal of replacing the winner [2] thus barely lines up with the way it is justified. It is impossible to agree with the subtle narrative that allegedly all voters who stopped supporting the administration candidate were not looking to change the winner and merely wanted to send the "party of power" a message.

It is obvious for us that the voters who voted for administration candidates in 2014, yet called off their support for them in 2019 were faced with a choice: which opposition candidate to vote for. It is likely they were faced with this choice even more so than those who voted for left-wing opposition in 2014. This is why the assistance of "Smart Voting" came just in time for them. As a result, many of these voters were engaged in both protest and tactical voting, which is why separating the two types seems forced.

Andrei Buzin's model [3] is also based on the differences in the 2014 and 2019 election results, and uses a unified formalized approach as well. This is why the same criticisms as in case with Ivan Bolshakov and Vladimir Perevalov can be made about this model. An attempt to engineer the overflow of votes from theoretical standpoint is what makes Buzin's study interesting. Andrei Buzin offers two alternatives: 1) distributing votes from the "party of power" and parties not contesting the election pro rata with results of the 2014 election; 2) consolidating all received votes with the opposition party that gained the most votes in the given constituency in the 2014 election. The actual outcome was close to the second alternative, which the author himself considers unrealistic. However, this is simply a demonstration of votes consolidating, not a confirmation. The main point here is that votes could not overflow following a single pattern, as the situation varied among constituencies depending on a number of factors.

Attempts to assess the effect of "Smart Voting"

This is the assessment made by Boris Ovchinnikov [12] immediately after the election: "The median result of CPRF candidates (excluding Vadim Cumin) not endorsed by "Smart Voting" amounted to 17%, whereas that of those endorsed by "Smart Voting" amounted to 35%". The difference in 18%. A Just Russia candidates follow a similar pattern: 7% not endorsed by "Smart Voting" endorsement (excluding the constituencies where pro-government communists ran in place of "inbearpendents") and 27% endorsed. +20%. In other words, a typical "Smart Voting" candidate received 18–20 percentage points more than his or her fellow party member without Navalny's endorsement."

This way, Ovchinnikov also addressed the cases when a CPRF or A Just Russia candidate did not have any competition among the "Smart Voting" endorsement contenders, which is not exactly on point. The existence of spoilers and liberal candidates, which could create distortions, was not addressed either.

To get a cleaner result, we decided to study only 14 constituencies, where candidate selection was the same (see section "The challenges of assessing the impact of "Smart Voting"). "Smart Voting" endorsed CPRF candidates in 11 constituencies. The average for CPRF candidates in these constituencies was 34.9%, the median – 35.1%. The average for A Just Russia candidates was 7.8%, the median –7.9%.

"Smart Voting" endorsed A Just Russia candidates in three constituencies. Their average in these constituencies was 25.2%, but the most remarkable thing is that all three results are very close – between 25.0 and 25.6%. CPRF's average in these constituencies is 17.8%, with median at 14.9%.

In other words, this sampling indicates an almost similar increment of nearly 17 percentage points. Can it be attributed to "Smart Voting"?

First, we need to see if these constituencies present any other electoral differences. The following are the average results of other groups of candidates: 38.8% (11) and 34.9% (3) for administrative independents, 7.7% (11) and 9.6% (3) for LDPR, 7.3% (11) and 8.1% (3) for Communists of Russia. We also looked up the average results in these constituencies in the 2016 legislative election: 13.8% (11) and 13.3% (3) for CPRF, 6.9% (11) and 5.7% (3) for A Just Russia. The difference is not critical, and in any case, the three constituencies where "Smart Voting" endorsed A Just Russia candidates were not exactly rewarding for them.

Therefore, it follows that "Smart Voting" candidates were more successful than their fellow party members without such endorsement were. Andrei Buzin [3] asks a reasonable question: what is the effect and what is the cause of such an unusual result? Did "Smart Voting" affect the result, or did its organizers just make a lucky guess about the most promising candidate? Andrei Buzin himself is inclined towards the first version, but does not give enough proof to such a conclusion. On the contrary, Ivan Bolshakov and Vladimir Perevalov [1; 2] indicate that a candidate's potential and chances are already built into the "Smart Voting" system as its basic criteria.

Let us try to look into this matter by using the aforementioned data on campaigning activity of various candidates [7]. Analysis of the six constituencies where "Smart Voting" settled on endorsing A Just Russia candidates while dismissing CPRF candidates implies that increased campaigning activity of the former, which was absent from five constituencies out of six (with the exception of Mikhail Timonov) was not the deciding factor – it was the low campaigning activity of the communists. As we mentioned before, A Just Russia candidates endorsed by "Smart Voting" – Vladimir Kalinin (no. 26), Sergei Vasilyev (no. 35) and Igor Glek (no. 38) – had sufficient financial resources, but their activity was barely visible. Ilya Lifantsev (no. 13) and Arkadii Pavlinov (no. 28) were neither visibly active, nor financially noteworthy. Nevertheless, the results of Lifantsev and Pavlinov are very similar to those of Kalinin, Vasilyev and Glek, and much higher than those of the more active and durable A Just Russia candidates Aleksandr Luchin (no. 11), Georgii Fedorov (no. 14), Boris Kagarlitsky (no. 42) and Aleksandr Romanovich (no. 37).

This brings us to the following conclusion: although campaigning activity played a role in choosing candidates for "Smart Voting" endorsement, this factor alone cannot explain the sizeable difference in the results of candidates who were and were not endorsed by Alexey Navalny's team.

The proposed model

Let us try to build our model from the assumption that every major party has its own stable (basic, core) electorate [14]. In light of that, most voters still make their decision during the campaign and support whichever candidate is the most active or whose campaigning they relate to the most.

Let us try to assess the percentage of each party's electorate among active voters based on the 2014, 2016 and 2019 vote returns. In this case, turnout differences are inessential.

CPRF candidates received an average of 20% in 2014. In 2016, CPRF's average in Moscow amounted to 14%. In 2019, the average of CPRF candidates not endorsed by "Smart Voting" amounted to 18%. However, even these candidates could receive additional votes from voters who decided on their choice during the campaign. For this reason, we are going to estimate CPRF's core electorate in Moscow at 14%.

A Just Russia candidates received an average of 10% in 2014. In 2016, A Just Russia's average amounted to 7%. In 2019, the average of A Just Russia candidates not endorsed by "Smart Voting" amounted to 8%. Therefore, 7% is quite a plausible estimate of this party's core electorate.

LDPR received 13% in Moscow in 2016, but a high proportion of this number is personal support of Vladimir Zhirinovsky. In 2014 and 2019, LDPR candidates received an average of 7% and 9% respectively. LDPR's core electorate in Moscow can be reasonably estimated at 7%. Note that voting for LDPR candidates in Moscow is protest from the political science standpoint, as its candidates typically have no chance of winning and they do not do much campaigning as a rule.

Candidates from RUDP Yabloko received an average of 12% in 2014. In 2016, RUDP Yabloko received 9.5% in Moscow while 2.6% more voted for People's Freedom Party (PARNAS) and 3.5% voted for Party of Growth. The sum total is 16%, but we are going to estimate the core liberal electorate at 12% (which amounts to the sum total of votes for Yabloko and PARNAS in 2016).

Unfortunately, not much can be said about core electorate of Communists of Russia. In 2016, this party received 2% in Moscow. In 2016, the average of its candidates amounted to 7%. The motivations behind the votes for this party are various: they include accidental voting, protest voting against CPRF's political behavior and attitudes along the lines of "I support communists, doesn't matter which ones."

United Russia's core electorate is the most difficult to estimate, as this party and its candidates usually receive an inflow of votes following mass campaigning. It is likely that the most reasonable estimate would stand at 18%, which is a result received by Vera Shevchenko in constituency no. 21, who did not qualify for victory, yet was registered as a secretary at a local United Russia branch office.

As a result, we estimate the aggregate core electorate of major parties at 58%. The remaining 42% are voters who decide on their choice during the campaign. Yet in the constituencies without liberal candidates (40 were like this), these deciding voters were joined by 12% of the liberal voters.

It seems between 3 and 29% of votes went to administration independents (with the exception of Yekaterina Kopeikina and Vera Shevchenko) at the expense of deciding voters, up to 32% went to CPRF (with the exception of Leonid Zyuganov and Nikolai Gubenko) and A Just Russia candidates each, and between 16 and 33% went to liberal candidates.

So, is it possible to assess which portion of these additional votes was received via candidate's personal campaigning activity and which was received via "Smart Voting"? Let us try to make such an assessment for the 33 constituency where the tactical voting choice was between CPRF and A Just Russia candidates first.

To eliminate the impact of various factors, it would pay to choose a small number of representative constituencies where both opposition candidates were mostly passive and their sum total results about the same. In our estimation, these are constituencies no. 10, 26, 38 and 39. The aggregate results of CPRF and A Just Russia candidates there varies between 38.1 and 39.8%.

"Smart Voting" endorsed CPRF candidates in constituencies no. 10 and 39 (Yurii Dashkov and Aleksandr Vidmanov). The former and the latter received 28.5% and 28.1% respectively. Compared to the 2016 CPRF results, their respective added votes amounted to 13.9 and 14.6%. These are the lowest results for CPRF candidates endorsed by Alexey Navalny's team (with the exception of Vladislav Zhukovsky). A Just Russia candidates in these constituencies (Andrei Suvorov and Andrei Bezryadov) received 10.3 and 10.6% respectively; added votes amounted to 3.6 and 6.0%.

"Smart Voting" endorsed A Just Russia candidates in constituencies no. 26 and 38 (Vladimir Kalinin and Igor Glek). They received 25.0% each. Compared to the 2016 results of A Just Russia, their respective added votes amounted to 19.3 and 20.8%. These results were also the lowest for A Just Russia candidates endorsed by "Smart Voting". CPRF candidates in these constituencies (Andrei Ispolatov and Lyudmila Yeryomina) received 13.1 and 14.9% respectively; added votes amounted to 0.7 and 1.6%.

We believe that the difference between constituencies no. 10 and 39 on the one hand and constituencies no 26 and 38 on the other encapsulates the effect of "Smart Voting", as the impact of the candidates' campaigning activity and the impact of any other candidates is nearly non-existent. If we take the difference between average results, it amounts to 14.3 and 14.6% for CPRF and A Just Russia respectively. If we take the difference between average added votes, it amounts to 13.2 and 15.2% for CPRF and A Just Russia respectively.

These estimates indicate that "Smart Voting" effect amounted to 14% for the cases where the choice was between CPRF and A Just Russia candidates. These are slightly less than the estimates made by Boris Ovchinnikov (between 18–20%).

Now, let us try to assess the "Smart Voting" effect on the results in the 20 constituencies with winning opposition candidates (Table 4).

Out of the 33 constituencies where CPRF and A Just Russia candidates ran against each other without any liberal candidates, they won 10 and 1 constituencies respectively. If we concentrate on our assessment of "Smart Voting" effect, we will be able to establish its role in these 11 constituencies.

Out conclusion is that "Smart Voting" had been decisive in A Just Russia's Mikhail Timonov winning in constituency no. 16 (he beat Anton Molev by a margin of 6%). The candidate certainly worked hard to win, but without endorsement from Alexey Navalny's team he would likely have failed to gain the necessary votes.

For CPRF, "Smart Voting" had been decisive in Dmitry Loktev winning in constituency no. 2 (2.7% margin), Pavel Tarasov in constituency no. 24 (6% margin) and Nikolai Zubrilin in constituency no. 11 (10% margin). Therefore, our conclusion is that "Smart Voting" had been decisive in three constituencies out of 10. In five more constituencies where communists beat administration candidates with A Just Russia candidates present (constituencies no. 17, 18, 20, 31, 42), CPRF candidates won by a margin of more than 14%, which is why it is reasonable to believe they would have won even without endorsement from Alexey Navalny's team. This is even more so concerning the two constituencies where "Smart Voting" endorsed candidates who were also endorsed by the mayor's office (no. 21 and 37).

Assessing the remaining 9 constituencies with winning opposition candidates proves to be more difficult.

In constituency no. 44, the "Smart Voting"-endorsed Yelena Shuvalova ran against the administration independent Ilya Sviridov, a member of A Just Russia. Essentially, voters also had to choose between CPRF and A Just Russia in this case, and Shuvalova won only by a 2.6% margin, so we assume that endorsement from Alexey Navalny's team had been decisive.

There were no A Just Russia candidates in constituencies no. 15 and 19, so the votes would have consolidated around CPRF candidates even without assistance from Alexey Navalny's team. Nevertheless, "Smart Voting" took effect here as well. Oleg Sheremetyev beat the administration candidate by a 2.0% margin, so the campaign's effect had definitely been decisive in this case. In constituency no. 15, Sergei Savostyanov won by a 10% margin, so here "Smart Voting" effect can mostly be assumed.

Following CPRF's withdrawal in constituencies no. 3 and 45, the votes consolidated around A Just Russia candidates Aleksandr Solovyov and Magomet Yandiyev respectively, although neither of them campaigned. However, an independent PARNAS member Mikhail Konev ran in constituency no. 45, but he brought discredit on himself by filing a claim against Ilya Yashin, which even led to him being expelled from the party. In constituency no. 3, Aleksandr Solovyov (who did not do any actual campaigning) beat the administrative independent Sabina Tsvetkova only by a 2% margin all while Communists of Russia candidate Leonid Voskresensky received 20.6% and LDPR's Yurii Shevchenko, who filed a case for removing the CPRF candidate Timur Abushayev from the election received only 7.2%. This case clearly indicates a general effect of tactical voting, but we may assume that "Smart Voting" campaign produced its effect as well, otherwise a part of the votes could go to Communists of Russia and LDPR. The same conclusion can be made for constituency no. 45. Magomet Yandiyev beat the administration independent Valeria Kasamara by a 5.7% margin while Mikhail Konev received 9.9% and LDPR's Yevgeny Turushev – 12.1%.

There was no administration candidate in constituency no. 43, so Sergei Mitrokhin's victory over LDPR's nominee by a margin of 24% was inevitable. In constituency no. 6, Yevgeny Bunimovich beat the administration independent Mikhail Balykhin by 13%; the effect of "Smart Voting" here is likely in the fact that it did not hamper RUDP Yabloko's candidate: if the stake had been on CPRF candidate (like in constituency no. 30), it would have been harder for Bunimovich to win.

In constituency no. 8, the administration independent Yelena Kopeikina served a technical purpose, so the real competition was going on between the communist Vadim Kumin endorsed by the city administration and the independent Daria Besedina, a member of RUDP Yabloko. She beat Kumin by 5%. We may assume that Besedina would have won anyway without any endorsement from "Smart Voting", but if Alexey Navalny had endorsed Vadim Kumin, the result would have been different.

Initially there was no CPRF candidate in constituency no. 14, but the party endorsed Sergei Tsukasov there. After registration, four opposition candidates appeared in the constituency at once: the independent Sergei Tsukasov, RUDP Yabloko's Maksim Kruglov, the independent Dmitry Klochkov (from Dmitry Gudkov's team) and A Just Russia's Georgii Fedorov, who was running an active campaign. As a result, there was an actual threat of opposition votes splitting. The removal of Tsukasov by a court order reduced the threat, but only a little. "Smart Voting" placed the stake on Kruglov, and so did Tsukasov. Kruglov subsequently gained 39.3% and beat the administration independent Natalia Pochinok by 16.5%. Fedorov and Klochkov received 13.1 and 5.2% respectively. As a result, the sum total of the opposition's results amounted to 58%, with two thirds received by RUDP Yabloko candidate. This is certainly a cumulative effect of the candidate's campaigning activity as well as the endorsement of Navalny and Tsukasov. At the same time, we cannot disregard the fact that endorsement from Tsukasov (who holds influence over leftist electorate) would have been enough, so in this case it is not necessarily the effect of "Smart Voting" as much as the idea of consolidating votes via tactical ("smart") voting.

That said, we can assert the decisive effect of "Smart Voting" in the opposition's victory in eight constituencies (no. 2, 3, 11, 16, 19, 24, 44 and 45) and assume it in three more (no. 8, 14 and 15). Let us note that Ivan Bolshakov and Vladimir Perevalov [2] recognized the campaign's effect in four constituencies (no. 2, 8, 19 and 44) and deemed four more (no. 3, 11, 24 and 45) as borderline.

On the other hand, we have to agree with Ivan Bolshakov and Vladimir Perevalov [2] in that the mistake of "Smart Voting" campaign led to the opposition's defeat in constituency no. 30, where the liberal independent Roman Yuneman narrowly missed the victory by a mere 0.25%. This is an example of how liberal voters fail to view reputable recommendations in favor of a communist candidate as a directive in the context of having to choose between a communist and a liberal candidate. It is nevertheless easy to assume that endorsement from "Smart Voting" would have enabled Yuneman to receive additional votes, as even 0.5% would have helped him win.

Ivan Bolshakov and Vladimir Perevalov assessed the choice of "Smart Voting" as misplaced in four more constituencies, where "Smart Voting" endorsed A Just Russia candidates. However, the context in these constituencies is different. For example, placing the stake on A Just Russia in constituencies no. 13 and 28 was indeed a mistake: their result was about the same as that of communists not endorsed by Alexey Navalny, and it is understandable that the latter would have received more had they been endorsed by "Smart Voting". However, if the communist Aleksandr Potapov had had a good chance of beating the administration candidate Igor Buskin, who had him by 6.6% in constituency no. 13, then the communist Konstantin Lazarev in constituency no. 28 would have required more than 15% to win, and it is unlikely that "Smart Voting" would have been able to ensure that. As for constituencies no. 26 and 38 where A Just Russia candidates beat CPRF by 12 and 10% respectively and fell behind administration independents by 12 and 11% respectively, one can be sure that Alexey Navalny placing the stake on communists would have failed to claim victory for the opposition.

Conclusion

The 2019 Moscow City Duma election certainly played a role in the development of electoral technology. "Smart voting" has become one of the main contributions. There is no doubt this technology will be used in the future, improving and adapting depending on the circumstances. With that said, political science has to learn how to assess the effect of "smart voting" on both vote returns and election results. We hope that our discussion will prove to be the first step in this direction.

Andrei Buzin [3] is adamant that Alexey Navalny's political strategists outmaneuvered their city administration colleagues. Multiple aspects of this assertion can be disputed, including the reasonable nature of the word "outmaneuvered". Two points are indisputable, however. First is the success of the "smart" (or tactical) voting technology: the opposition would have gained much less seats in Moscow City Duma (less than half by our estimates). Second, the nature of this success was technological rather than political.

The fact that this successful application of political science was inspired by the changed political circumstances is a different matter. It is unlikely that the same result could have been achieved in 2014, when the "party of power" was gaining momentum. Political situation of 2019 was entirely different, yet it required skillful handling in the context of city administration still counting on dominating the City Duma.

That said, we have to admit that political strategists in the mayor's office made a series of miscalculations. Saying their skill is as sharp as ever would likely be an overstatement. The press pointed out that election managers in the mayor's office were replaced, so the 2019 election was handled by people without proper experience [13].

The question of how denying the liberal opposition candidates participation in the election affected the results remains open in particular. Successful campaign of the three RUDP Yabloko candidates and Daria Besedina along with Roman Yuneman's relatively successful performance suggests that Gennady and Dmitry Gudkovs, Yulia Galyamina and Ilya Yashin had high odds of winning. At the same time, registering a host of other candidates could divide the opposition votes between stronger candidates: Ivan Zhdanov and Daria Besedina in constituency no. 8, Andrei Babushkin and Nikolai Zubrilin in constituency no. 11, Konstantin Yankauskas and Lyubov Nikitina in constituency no. 31, Anastasia Bryukhanova and Yekaterina Yengalycheva in constituency no. 42, Lyubov Sobol and Sergei Mitrokhin in constituency no. 43. Besides, removing candidates from the election in court cases (Timur Abushayev in no. 3 and Sergei Tsukasov in no. 14) as well as forced withdrawal of Vladislav Kolmagorov in no. 45 clearly contributed to the decreased splitting of opposition votes. As a result, the extra diligence of administrators and political engineers from the city administration contributed to the defeat of its candidates in some cases.

Another noteworthy aspect is related to the choice of electoral system. The legislators' gravitation towards trying to bring back the plurality voting system whenever possible or to decrease its share in the mixed system has become a recent trend. Switching the Moscow City Duma election to plurality voting in 2014 was the first example of its kind, which was followed by similar decisions in a host of regions and cities. These decisions were motivated by a desire to maintain United Russia's dominance in the legislature amidst its decreasing popularity [6].

However, plurality voting is only lucrative for the leading party under certain conditions. One of these conditions is maintaining a high enough level of support. Success is almost guaranteed at 40%, very likely at 35%, whereas at a less than 30% level of support chances of success are unlikely. A graphic illustration of this fact can be found in a series of elections held in 2018–2019 (in the cities of Tolyatti and Dimitrovgrad, as well as in Khabarovsk Krai, including the cities of Khabarovsk and Komsomolsk-on-Amur) [6; 5].

The second condition is a lack of consolidation of other parties against the leading party. United Russia maintained the image of a centrist party for a long time, which allowed it to contest both the left and the right. However, consolidation of leftist and rightist politicians against the "party of power" has been gradually gaining momentum lately, facilitated by United Russia's anti-democratic and anti-social policy. The "Smart Voting" campaign is one of the results of such consolidation. It is likely the campaign will stifle the trend of switching elections to the plurality voting tracks among other things.

Received 25.05.2020, revision received 01.06.2020.

---

References

1. Bolshakov I.V., Perevalov V.V. Assessing the effectiveness of "Smart Voting" strategy: a discussion of analytical approaches. – Electoral Politics. 2020. No. 1 (3). - http://electoralpolitics.org/en/articles/otsenka-effektivnosti-umnogo-golosovaniia-spor-analiticheskikh-podkhodov/
2. Bolshakov I.V., Perevalov V.V. Konsolidatsiya ili protest? "Umnoye golosovaniye" na moskovskikh vyborakh [Consolidation or protest? "Smart Voting" in Moscow election]. – Politeia. 2020. No. 1 (96). P. 50–73. DOI:10.30570/2078-5089-2020-96-1-50-73. (In Russ.)
3. Buzin A.Yu. Someone else's victory. – Electoral Politics. 2020. No. 1 (3). - http://electoralpolitics.org/en/articles/chuzhaia-pobeda/
4. Kolosov V.A., Vendina O.I. Politicheskiye predpochteniya moskvichei v khode izbiratelnykh kampanii [Political preferences of Moscow citizens during election campaigns]. – Herald of the Russian Academy of Sciences. 1997. Vol. 67. No. 8. P. 675–680. (In Russ.)
5. Kynev A.V., Lyubarev A.E. Itogi golosovaniya i rezultaty vyborov 8 sentyabrya 2019 goda. Analiticheskiy doklad No. 6 po Monitoringu vyborov 08.09.2019 [Vote returns and election results on 8 September 2019. Analytical report No. 6 on monitoring activity of the September 8, 2019 election]. – "Liberalnaya missiya" foundation website, 2 October 2019. URL: http://liberal.ru/files/articles/7409/Vibory_2019.Itogi.pdf (accessed 24 May 2020). (In Russ.) - http://liberal.ru/files/articles/7409/Vibory_2019.Itogi.pdf
6. Kynev A.V., Lyubarev A.E., Maksimov A.N. Pravovyye i politicheskiye osobennosti vyborov 8 sentyabrya 2019 goda: rost departizatsii i personalizatsii vyborov. Analiticheskiy doklad No. 1 po dolgosrochnomu nablyudeniyu vyborov 08.09.2019 g. [Legal and political characteristics of the elections on 8 September 2019: growing de-partisation and personalization of the election. Analytical report No. 1 on monitoring activity of the September 8, 2019 election]. – "Liberalnaya missiya" foundation website, 7 July 2019. URL: http://liberal.ru/files/articles/7388/Vibory_2019.Analiticheskiy_Doklad_1.pdf (accessed 24 May 2020). (In Russ.) - http://liberal.ru/files/articles/7388/Vibory_2019.Analiticheskiy_Doklad_1.pdf
7. Kynev A.V., Pozhidayeva E.O., Khlopov A.V. Osobennosti kampaniy po okrugam [Campaign specifics constituency-wise]. – Transregional Voters' Association website, 31 May 2020. URL: http://www.votas.ru/book-Moscow/7-2.html (accessed 01 June 2020). (In Russ.) - http://www.votas.ru/book-Moscow/7-2.html
8. Lyubarev A.E. Analysing the results of the 2019 Moscow City Duma election: a preface to the discussion. – Electoral Politics. 2020. No. 1 (3). - http://electoralpolitics.org/en/articles/analiz-rezultatov-vyborov-v-moskovskuiu-gorodskuiu-dumu-2019-goda-predislovie-k-diskussii/
9. Lyubarev A.E. Raiony Moskvy na vyborakh 2012–2018 godov [The districts of Moscow in the 2012–2018 elections]. – Transregional Voters' Association website, 31 March 2018. URL: http://www.votas.ru/Moscow-12-18.html (accessed 24 May 2020). (In Russ.) - http://www.votas.ru/Moscow-12-18.html
10. Lyubarev A.E. Vybory v Moskve: opyt dvenadtsati let. 1989–2000 [Elections in Moscow: a 12-year legacy. 1989–2000]. Moscow: Stolny grad, 2001. 416 p. (In Russ.)
11. Ovchinnikov B. Effekt spoilerov: kak "Kommunisty Rossii" podarili vlasti ot 8 do 13 mandatov v Mosgordume [Spoiler effect: how Communists of Russia presented the authorities with 8 to 13 seats in Moscow City Duma]. – Golos Movement website, 26 September 2019. URL: https://www.golosinfo.org/articles/143791 (accessed 24 May 2020). (In Russ.) - https://www.golosinfo.org/articles/143791
12. Ovchinnikov B. Effekt "umnogo golosovaniya" v Moskve: 300 tysyach golosov [The effect of "Smart Voting" in Moscow: 300 thousand votes]. – Golos Movement website, 2 October 2019. URL: https://www.golosinfo.org/articles/143799 (accessed 24 May 2020). (In Russ.) - https://www.golosinfo.org/articles/143799
13. Pertsev A. Metaniye tortami drug v druga. Kto vinovat v tom, chto v Moskve razrazilsya politichesky krizis I pochemu meriya okazalas k etomu ne gotova [Throwing cakes at each other. Who is to blame for the political crisis in Moscow and why the city administration was not prepared]. – Meduza. 26 July 2019. URL: https://meduza.io/feature/2019/07/26/moskovskiy-krizis (accessed 24 May 2020). (In Russ.) - https://meduza.io/feature/2019/07/26/moskovskiy-krizis
14. Podvintsev O.B. Elektorat [The electorate]. – Vybory i elektoralnaya politika: slovar. SPb., 2010. P. 154. (In Russ.)