The article discusses the approaches for assessing the impact of "Smart Voting" strategy on the 2019 Moscow City Duma election results. The authors use electoral behavior theory to demonstrate how the change of public opinion, protest mobilization and electoral space transformation ushered a victory for opposition candidates. Electoral behavior of Moscow citizens has changed significantly compared to the last election: there is an obvious increase of protest and tactical voting. These voting methods are different by nature, motives and goals, but this time they found common ground in candidates, making it possible for analysts to discuss the effectiveness of the opposition's new strategy. The authors insist that "Smart Voting" should be regarded as tactical voting, as protest voting has no direct relationship with this strategy. Measuring the quantitative impact of the two voting methods helps to identify no more than six constituencies where the strategy was most significant and five constituencies where it could have been significant had other candidates ended up on the "Smart Voting" list. However, the strongest impact on election results was created by a decreased competitiveness across constituencies and the change of public opinion that allowed focusing voters' attention on the more consequential opposition candidates.
Many reporters and analysts called the results of September 2019 election to Moscow City Duma sensational. For the first time in many years, the opposition challenged the authorities and nearly stripped them of parliamentary majority. Representatives from opposition parties took 20 out of the 45 seats, which only compares to the structure of the first two convocations of the city parliament. This achievement is attributed to "Smart Voting" strategy initiated by Alexey Navalny. The essence of this strategy is to vote for the candidate who is the strongest alternative to the authorities.
In the context of electoral authoritarianism, these election results are indeed impressive, although the optimistic conclusions regarding the effectiveness of Navalny's strategy are unsubstantiated and premature. Our paper on assessing the impact of "Smart Voting" on election results published in the journal "Politeia"  gave the discussion of the strategy a new turn. For this reason, we find that we have to clarify our position on the issue and to elaborate as to why our model provides a better interpretation of Moscow City Duma election results than its alternatives.
When it comes to Moscow elections, researchers focus on the list of candidates endorsed by "Smart Voting" while paying no heed to the multitude of factors, voters' motivations and causal links. The judgement on the strategy's success is made intuitively [29; 6; 28] while the quantitative assessment is reduced to pure mathematics [25; 5] and isolated from the context and electoral behavior theory. Analysts overlook the "Smart Voting" itself and every time view it as something else instead: a protest, tactical voting, Navalny's influence or everything at once.
For example, Kirill Rogov uses protest and tactical voting as synonyms. For this reason, the gains in electoral indicators of "Smart Voting" candidates (as compared to electoral indicators of previous election's runner-ups) are attributed to the strategy itself while the connection with other candidates' results in the given constituencies is subsequently ignored.
In his paper, Andrei Buzin measures how the results of party candidates has changed compared to the 2014 election. The study shows that the results changed significantly if the candidates were endorsed by "Smart Voting," and either dropped or slightly changed if they were not. Although this analysis is helpful in measuring the dynamics of party preferences, it does not help us understand the Moscow City Duma election results. The results do not depend on party preferences as much as they do on voters' motivations and structural characteristics of electoral space. At the same time, it is clear that both the public mood and the structure of electoral menu have shifted dramatically compared to the previous election.
For his part, Boris Ovchinnikov compares the results of the Communist Party of the Russian Federation (CPRF) candidates who were endorsed by "Smart Voting" and of those who were not to find that the "Smart Voting" candidate gained 18-20 pp more votes than their fellow party member not endorsed by Navalny. That said, the conclusion made from analysing the indicators of one party is extrapolated to all constituencies. Not only does this method of qualitative assessment prevent any accurate measurements from being recorded, but it also generalizes special cases, which naturally leads to data corruption.
Buzin and Ovchinnikov fail to take one major factor into account: a candidate's potential and chances are already built into the "Smart Voting" system as its basic criteria, meaning they already serve as an explanatory variable to both candidate's nomination for "Smart Voting" and election results. By studying the link between being on "Navalny's list" and vote returns we combine two dependent variables while breaking the principle of consequential events.
In reality, Buzin and Ovchinnikov compare the results of candidates who, in many cases, were one of the few, if not the only ones among the opposition to run a campaign in the constituency, or were the only significant challenger to the party of power, to the candidates who either did not run a campaign or were overshadowed by the more powerful and active members of the opposition. In the end, both Buzin and Ovchinnikov assume that the percent interval between these two groups of candidates is linked to the "Smart Voting" impact. The findings from this model lose any meaning, as the link between the given variables is mediated by a third variable, which is removed from the analysis itself. This approach is justified if tactical voting is employed accidentally and heuristically, when voters themselves evaluate the chances of a candidate. In our case, however, such approach proves to be invalid.
Moreover, the 2019 election results do not just involve the increased support for opposition candidates; the point is that they have surpassed their pro-government competitors, none of whom went over 50%. In the previous election, 17 candidates from the party of power gained the absolute majority of votes (the highest result being 69.28%). In the current election, only 12 administrative candidates garnered the support of more than 40% of voters (the highest result being 47.17%). In light of this, we cannot disregard the decreased support for administrative candidates and confine all changed voter preferences to the opposition electorate alone. However, interpreting all additional votes gained by alternative candidates as the effect of "Smart Voting" appears unjustified without at least identifying the strategy's mechanism for influencing voter motivation and putting the object of measurement into the collection of variables that can be empirically tested.
Electoral studies cannot be conducted without mathematical methods just as much as they cannot be conducted without logic and defining the studied phenomena in the context of political science and sociology. Interpreting election results without any theoretical basis and context consideration leads to what Giovanni Sartori calls "conceptual stretching" . In other words, by disregarding object pre-identification and qualitative evaluation of concepts we engage in false measurement, which does not quite answer the question we try to address through quantitative analysis. This is why we cannot consider either Rogov's (200 thousand or 12.5% of votes) or Ovchinnikov's (300 thousand or 19% of votes) estimates a valid measure for "Smart Voting" strategy.
We believe that "Smart Voting" has to be identified just as its authors envisioned it – a way to consolidate the electorate to prevent the votes of the opposition supporters from scattering. The following is the quote from the authors of the strategy: "We need 3% of those voters on the general electoral roll who never went to the polls before to take part in "Smart Voting" as well as the third of the voters who always goes to the polls to vote against United Russia. 33% not from the general electoral roll, but from those who always voted opposition. It's just that this time they have to vote for the same candidate. Then we'll definitely take the majority away from United Russia. What's important is that they have to vote smart instead of voting ideologically, for 'their' party" . "In a real election, everyone votes on their convictions, not tactically. But there's little hope for that after everything that we've seen recently. As a result, they pretend it's a real election and so do we. But even in a situation like ours we can try to maximize the damage to the government" .
These statements indicate a typical tactical voting scenario where the voters make their vote as useful as possible: they do not waste it by voting for underdog candidates and instead choose someone with the best chance of defeating the disliked candidate that is expected to win . The motives behind this most pragmatic type of electoral behavior have nothing in common with personal preferences determined by the personality of a candidate, agreeing with their political platform (ideological voting) or a desire to demonstrate dissatisfaction (protest voting). This means that a distinction must be made between tactical and protest effects of voting, and that their impact on election results must be measured separately, which we did in our study.
We divided the election participants into three groups: administrative candidates endorsed by the mayor's office, alternative candidates endorsed by the "Smart Voting" list, and the third group of candidates (all the remaining election participants who are not part of the main competition). The analysis also excluded atypical situations (constituencies no. 21 and 37), where administrative and alternative candidates coincided, meaning they were endorsed both by "Smart Voting" and mayor's office.
To solve the main problem, we measured the dynamics of electoral support for the three groups of candidates and the overflow of votes between them by comparing it with similar indicators from the 2014 election. (In 2014, the alternative candidates were the stronger opposition nominees who gained the most votes than other opposition candidates).
If the indicators of alternative candidates increased at the expense of the third group's indicators decreasing, then a tactical effect was produced (32 constituencies). If they increased at the expense of administrative candidates' indicators decreasing, this was a protest effect (35 constituencies). Then, by clearing the results of alternative candidates from one of the effects, we determined which of the effects was the source of the advantage over the administrative opponent and helped "Smart Voting" candidates win.
Tactical voting had a decisive impact only in 4 out of the 18 constituencies won over by the opposition: without it, alternative candidates were losing to administrative candidates by 3.2 to 13.1 pp. In 10 constituencies, alternative candidates won by a margin of 3.5 to 23.1 pp without tactical voting; in 6 out of these 10 constituencies the margin was over 15 pp. In four more constituencies, the effects were limited (between 0.43 to 0.85 pp for two cases of tactical voting and between 0.44 and 1 pp for two cases of protest voting). In 30 out of 43 constituencies, the decrease in support for administrative candidates exceeded the decrease in third group's indicators. The situation was the opposite only in 13 constituencies. Eleven cases indicated no tactical voting effect whatsoever – the third group gained more votes than in the previous election.
An analysis of the overflows revealed a tactical voting share of 5.57%. The share of protest voting was 11.13%: 9.72% for alternative candidates and a 1.41% overflow to the third group. This estimate corresponds with the scope of tactical voting in the countries where it is typically observed, namely the United Kingdom, Germany and Canada, its share amounting to no more than 6-8% [11; 13; 8; 14].
The correlation analysis confirms our estimates. However, individual comparison of the indicators in constituencies with an alternative (18 constituencies) and administrative (25 constituencies) majority allows for a more reliable conclusion. Constituencies with an alternative majority indicate a heavy dependence of "Smart Voting" candidates' support on decreased support of pro-government candidates (\(R\) = –0.77; \(R^2\) = 0.59). A direct link between the results of alternative and third group candidates (\(R\) = 0.24) is revealed here as well. This is indicative of similar patterns in supporting the "Smart Voting" candidates and minor opposition candidates, which should not be present under the conditions of tactical voting and votes overflowing from minor candidates. Constituencies with an administrative majority indicate a moderate dependence of the increase in alternative candidates' indicators on the decrease in third-group candidates' indicators (\(R\) = –0.57; \(R^2\) = 0.32). Despite a more significant impact of "Smart Voting" in this case, it was not enough for the opposition to win.
Another factor that should be considered when assessing the effectiveness of "Smart Voting" is the configuration of electoral space. Although, according to Arkadi Lyubarev’s estimates, the effective number of parties decreased only slightly (3.24 against 3.36 in the previous election) , comparing the structure of election participants reveals a link between election outcomes and party exposure. The index of the effective number of parties demonstrates the results of the competition, but does not reveal anything about the disposition of political actors prior to polling day.
In the 2014 election, all major opposition parties competed among themselves: candidates from CPRF, Yabloko, A Just Russia and the Liberal Democratic Party of Russia (LDPR) were running in 42 constituencies out of 45. In 10 constituencies, the competition was intensified by the presence of strong independent candidates or representatives of the Civic Platform party. One opposition party or another was not represented only in three constituencies. In four constituencies, United Russia did not nominate anyone, so the opposition nominees became administrative candidates there.
Within the opposition itself, CPRF and Yabloko became the main competing parties. The candidates from these parties went toe-to-toe in 19 constituencies at once: CPRF was the strongest opposition party with Yabloko as its runner-up in 16 constituencies while a reversed situation developed in 3 constituencies. 10 constituencies saw CPRF challenging A Just Russia, with 6 consituencies dominated by communists and 4 by A Just Russia. CPRF bested other opposition candidates in 9 more constituencies while Yabloko did the same in 3 more. A Just Russia did not top any other constituency and was the runner-up to the opposition candidate only once more. LDPR candidates became runner-ups only in three constituencies and bested the rest of the opposition in a single constituency (granted the nominee was the administrative candidate, this case should be considered an exception).
Civic Platform gained the upper hand in two constituencies (where it separately competed against CPRF and A Just Russia) and became the runner-up to CPRF in another. In seven constituencies, the race was going on between the opposition party and an alternative independent candidate. Five of these constituencies were dominated by communists, one by Yabloko and one by an independent candidate.
The year of 2019 saw a completely different balance of power. Nearly thirty alternative candidates representing Yabloko, Alexey Navalny, Dmitry Gudkov and other factions were either denied registration or removed from the election in a court case, or dropped out of the race for other reasons. United Russia did not officially nominate its candidates, so all representatives of the party of power ran as independent candidates (we will further refer to them as United Russia candidates for clarity nonetheless). At the same time, the number of administrative candidates representing the opposition parties increased to six. As a result, none of the constituencies saw all major parties compete against each other.
In 35 constituencies, CPRF, LDPR and A Just Russia candidates competed without anyone from Yabloko (United Russia candidates were not present in 4 of these constituencies either). The competition turned out to be even more limited in the remaining constituencies. Two of these constituencies had A Just Russia running against LDPR, but not Yabloko and CPRF. Two more constituencies where Yabloko was running against LDPR did not feature either CPRF or A Just Russia. In four cases, CPRF and LDPR were featured while A Just Russia and Yabloko were not. The main opposition parties (three parliamentary plus Yabloko) were simultaneously represented in two constituencies only, although they were only competing among themselves, as one of the opposition parties served as the administrative actor.
While 30 independent candidates ran in the previous election, this time there were only 16. In 2014, five independent candidates gained over 10% (22.99% max). In 2019, there were only two such independents (28.36% max). Voters supported the independent candidates more commonly even if they were a minor alternative: without considering the stronger independents, the mean values were 4.82% and 6.09% respectively.
As a result, the 20 winning "Smart Voting" candidates were not competing against Yabloko in 16 cases, against United Russia in 6 cases, against CPRF in 3 cases and against A Just Russia in two cases. Speaking of all constituencies, in 10 cases the election was held without the participants who previously were alternative candidates and gained an average of 22.5% votes while in 22 cases there were no participants serving as the second alternative candidate with an average of 12.6% votes. Considering the nature of electoral geography, the oppositional capacity of a constituency remained unchanged, but other participants profited from it this time.
An equally important factor for the 2019 campaign was the spoiler effect, which was not absent in 2014. CPRF had to compete against Communists of Russia in 30 constituencies, where it lost between 2.47 and 11.45 pp (6.66 pp on average) to the second communist party. The opposition lost 22 constituencies with Communists of Russia as a participant, while winning 9 (Communists of Russia did not run against CPRF in one of these). CPRF could have won 14 more constituencies had the spoilers not participated. It is impossible to make a definitive conclusion on alternative outcomes, as voters also relied on personal qualities of a candidate (for example, spoiler was voted for more often if it was the only female candidate or the only female opposition candidate). However, this victory was possible in at least three constituencies where the gap between the pro-government candidate and CPRF candidate was minimal (constituency no. 7 – 0.38 pp, constituency no. 32 – 1.32 pp, constituency no. 36 – 0.08 pp).
If the four parliamentary parties and Communists of Russia were simultaneously running in one constituency (meaning the major opposition party was running against both United Russia and no less than three other opposition parties, including the spoiler party), the opposition mostly lost (20 cases) and rarely won (4 cases). In any case, there were five losing cases in less competitive settings. In three constituencies, CPRF was running against United Russia, LDPR and A Just Russia, but without Communists of Russia while in two it was running against United Russia, LDPR and Communists of Russia, but without A Just Russia (without Yabloko in both cases as well).
The remaining 16 cases are winning cases, four out of which indicated little competitiveness (out of the main parties, one constituency only saw United Russia, A Just Russia and LDPR while the rest only saw CPRF, LDPR and A Just Russia). In three cases, only the four parliamentary parties were competing among themselves (without Yabloko and United Russia in all seven cases). Three more constituencies saw three parliamentary parties and Yabloko running against each other (no United Russia and Communists of Russia in the two and no CPRF and Communists of Russia in one). Another three had three parliamentary parties and Communists of Russia (no A Just Russia in two cases and no CPRF in one; no Yabloko in all three either). Finally, the remaining three constituencies had a relatively high competition, but there was always a party lacking in every case for full representation (Yabloko in two cases and one for A Just Russia and United Russia each).
Relatively high results achieved by Communists of Russia in several constituencies (higher than the average 7.11% in 14 cases, with 20% as the highest in a constituency without CPRF) are an indirect indication of protest and ideological voting. Voters guided by the "Smart Voting" list would not have mistaken one communist party for another, especially if it were a matter of a candidate's personality.
Such transformation of electoral space puts considerable strain on trying to conduct a pure analysis of "Smart Voting" impact. The authorities had already done what the new strategy was intended to do by decreasing competition levels among the opposition parties in some cases and releasing them from competition with party of power in other. A consistent opposition voter did not require a special tool for transferring their vote from a weaker candidate to a stronger one – many simply did not have any other choice except voting for the more prominent alternative candidate. This is equally applicable to protest electorate for which the party it votes for is not as significant as the act of protest. For these reasons, comparing party results is meaningless. This is why our analysis focuses on the indicators of administrative and alternative candidates.
To reduce the impact on the findings of the decreased competition factor, the effectiveness of "Smart Voting" should be measured where it was truly meaningful, namely in competitive constituencies where at least two alternative candidates could compete for the opposition majority. For this purpose, we selected the constituencies that complied with two conditions: a "Smart Voting" candidate competed against an administrative candidate from the party of power while also running against the second alternative opposition candidate. The second condition was defined as nothing more than a twofold gap between the candidates. We found six eligible constituencies (no. 3, no. 13, no. 26, no. 28, no. 30 and no. 38), where the first alternative candidate (also known as the "Smart Voting" candidate) averaged 26.87% and the second averaged 20.95%. In the remaining 33 constituencies, the alternative candidate was nearly unchallenged (36.80% against 10.11%). "Smart Voting" effect alone cannot explain the extremely low result of the runner-up.
Two more cases can be referred to as competitive with certain reservations, where there may have been a twofold gap between the results of alternative candidates, but not as wide as in other cases. In constituency no. 16, Mikhail Timonov, a candidate from A Just Russia, gained 36.40% (winning the election) while his CPRF contender, Aleksandra Andreyeva, gained 17.34%. At the same time, not only did she face an ideological spoiler in Communists of Russia (4.34%), but also an independent candidate of the same name (3.82%). In constituency no. 35, A Just Russia's Sergei Vasilyev gained the support of 28.36% voters (losing the election) while his CPRF runner-up Dmitry Agranovsky gained 11.83% votes (the candidate from Communists of Russia gained 7.97%).
That said, out of the first group of constituencies, only in constituency no. 3 did a "Smart Voting" candidate Aleksandr Solovyov outrun the administrative candidate and the Communists of Russia candidate, with the latter gaining the maximum of 20.62% there. Given that Aleksandr S. Solovyov was nominated by A Just Russia specifically to run against Aleksandr Yu. Solovyov, his independent namesake, who ultimately was not registered, and that CPRF's Timur Abyshev was removed from the election in a court case, the opposition competition took place between two spoiler participants. The remaining five constituencies are among those where "Smart Voting" failed by supporting less promising candidates. These include the extensively covered Roman Yuneman case as well as the constituencies where "Smart Voting" endorsed A Just Russia nominees while underestimating the CPRF candidates. That said, Vladimir Kalinin, a "Smart Voting" candidate, did not simply lose the election to the administrative candidate in constituency no. 26 – he also gained 7 pp less than the alternative candidate in the previous election, whereas the results of the third-group candidates increased by 21.4 pp.
We assume that candidate disqualification and protest voting that is often mistaken for "Smart Voting" became the main success factors for the opposition in the Moscow City Duma election. Although at a first glance it might seem that we are referring to a unified group of voters, or one and the same behavioral pattern where a vote is considered useful by the voter and they strive to bring down the results of pro-government candidates, we insist that "Smart Voting" is to be viewed as tactical, and that protest voting took place by itself, without being triggered by the strategy.
The logic of a protest vote is different from the model of preventing wasted votes. Tactical voters renounce the preferred candidate because he or she has no chance of winning the election. Protest voters turn their backs on the preferred candidate because they want to teach him or her a lesson. The behavior of both groups is pragmatic. However, the first group is driven by concern that the least preferred candidate is going to win, but motivated by the chance to avoid the undesirable outcome. The second group is driven by disaffection, but motivated by hope that the signal of disaffection will bring change. Tactical voters want to replace the main actor while protest voters want him or her to adjust his or her policy.
Since protest voters count on provoking the authorities into reacting to the lost votes (expecting that they will want to bring back the old supporters), they strive to make their voice truly heard. Having a viable alternative that will make the signal visible means protest voting is more likely to take place [15: 397]. In this regard, both tactical and protest voters may have the same candidate of choice.
In its essence, protest electoral behavior is a vote against the main parties with access to government formation processes and procedures. Since only one party has this kind of access in Russia, it is a matter of voting for any opposition party that the voter deems viable (the voting model in the 2011 Russian legislative election). This is why parties like CPRF, Yabloko and A Just Russia, which are more familiar to the region, were able to use disaffection to their advantage in the Moscow election while a party like LDPR, for example, was not, as it has never had any future in the region.
The examples of Aleksandr Solovyov and Magomed Yandiyev, who won their seats without any campaigning, often act as arguments in favour of "Smart Voting" being effective. This impression was further strengthened by Solovyov's disappearance immediately after the election, which made many people doubt his actual existence. However, protest support may sometimes be given to minor and even phantom parties that exist to receive the anti-elite protest vote [3; 26]. The second thing to remember is that both Solovyov and Yandiyev represented A Just Russia – a parliamentary party – and competed for protest votes against LDPR and Communists of Russia instead of Yabloko and CPRF, meaning voters perceived them as the main viable opposition candidates.
The outcomes of economic and protest voting are more obvious if there are just a few viable alternatives with a relatively small number of effective parties in the election [1: 155-156], but these outcomes are less likely " if there are no credible actors who can benefit from such behaviour" [2: 753]. The chances of success may be slim, but such candidates are the ones to attract the support of dissatisfied voters when there are no viable alternatives. Many regional elections in Russia follow this exact scenario. Despite purging the political field of any viable opposition, Kremlin-endorsed candidates lost the 2018 gubernatorial election to their validated alternates who never even intended to win . In March 2019, an LDPR-endorsed 28-year-old housewife beat the chairman of the city duma in the mayoral snap election in Ust-Ilimsk of Irkutsk Oblast. People voted for ordinary workers and intellectuals in the first alternative legislative election held in 1989 for the same reasons by setting them off against the Soviet nomenklatura (the elites).
In such context, Khabarovsk Krai legislative assembly election is telling as well, where LDPR candidates endorsed by "Smart Voting" won 23 out of 24 constituencies. Viktor Fedoreyev (a member of Yabloko running as an independent candidate), who was the only alternative winner of plurality-voting part of the election, did not run an active campaign and was not endorsed by "Smart Voting," but still won by a majority due to protest voting and the existing situation in his constituency. We must bear in mind, however, that Khabarovsk election was not defined by disaffection alone, but also by the choice between two administrative parties – United Russia and LDPR – which the latter became after its candidate won the gubernatorial election the year before. It is therefore obvious that protest voting often forwards the votes to minor candidates while tactical voting always does the opposite.
Nevertheless, "Smart Voting" may have resulted in unexpected added effects, such as an increase in protest mobilization and using the list of candidates as guidelines for protest voters. This hypothesis has the right to exist, although it would be impossible to confirm it without conducting a sociological study of voter motivations. Although it is possible to assume this impact based on indirect data, one should refrain from making any quantitative assessment. Besides, the focus shifts from the model of "Smart Voting" itself to the impact of having opinion leaders as political support. In the meantime, Navalny's list was not the only recommendation mechanism in this election. Mikhail Khodorkovsky's made his own list, as did Dmitry Baranovsky, a municipal deputy – the latter was quite popular among activists. In certain constituencies, candidates, local deputies and action groups formed local coalitions as well. This provided extra support for candidates, although their opinions and actions played a certain role when it came to garnering support from local coalitions while "Smart Voting" tried to eliminate these factors.
The bigger point is, however, that "Smart Voting" was intended for the opposition voters instead of protest voters. The first group are the staunch opponents of the government who vote opposition every time while the second group are the resentful and disaffected ex-loyalists who found themselves as temporary travelling companions to the opposition. As the joint study by Higher School of Economics and Indiana University covering protest mobilization in 2011-2012 indicates that although both loyalist and opposition voters may share opinions on economic policy and the government's work, their identities and understanding of political reality are different, they use different communication channels and consume information differently . The situation may have changed over the past years, but we neither have the data that would confirm it, nor the explanation for such change.
We cannot rule out the fact that a certain number of voters who decided to protest-vote looked to the chances of candidates in an attempt to assess the viability of the alternative and therefore knew perfectly well which candidates from their constituency were on the "Smart Voting" list. Still, even in this case, the main reason (protest) was more of a driving force behind the change of preference than the secondary reason (voting for a viable candidate), especially since the viability of alternative was obvious in most cases.
Denis Volkov and Andrei Kolesnikov write that Russians "do not focus on what has to be done to bring positive change and how to do it" and mostly mention what "kind of end result they would like to see instead of how to achieve it" . In this context, the greater awareness of a typical protest voter, his or her ability to understand political scenarios and the nuanced instrumentality of choice is questionable. Therefore, we assume that protest voters are a completely different audience and would hardly be able to follow Navalny's recommendations.
We are willing to concede that "Smart Voting" facilitated the mobilization of opposition voters (especially those who vote sporadically and whose participation in the election was up in the air after some visible candidates were denied registration) and helped to channel the protest energy towards active participation in the election. However, protest mobilization in itself cannot is not limited by the new strategy. At the very least, we were unable to register any direct correlation between the increased voter turnout and supporting alternative candidates. Quite the contrary: wherever turnout was increasing, "Smart Voting" candidates received less votes.
General turnout indicators paint the same picture (an average increase of 1.2 pp in constituencies with alternative majority and 0.3 pp in constituencies with administrative majority). Then again, Ovchinnikov argues that the electorate structure itself changed in the 2019 election. By juxtaposing turnout geography with party preference geography in the 2016 election, he infers that pro-opposition voters outnumbered pro-government voters . However, his method of assessing voter preferences does not just fail to describe their behavior this particular time; it also works only if people identify themselves with the parties they vote for, which is not the case in reality. Sociologists from the Russian Public Opinion Research Center (VTsIOM) did a curious study that proves that "a large percentage of respondents easily renounce their opinion and change their declared political stance" [19: 33]. Over several survey ways, 38% of respondents changed the party they were planning to vote for. At the same time, 70% of this group responded negatively to the question of whether they changed their choice. Moreover, between 20% and 32% of respondents changed their retrospective answer over just 2.5 weeks, meaning that at least twice they mentioned a different party as the one they voted for in the previous election.
Ovchinnikov's model demonstrates that the structure of the electorate have changed by 9 pp compared to 2016, while compared to 2014 the change was twice as big. Considering the electoral volatility range in Moscow (about 14% in the 2011-2016 federal elections and about 16% in the 2009 regional election by the Pedersen index), the acquired data does not indicate any radical change in the structure of the electorate. Besides, it is difficult to see from this model how exactly these changes are stimulated by "Smart Voting" and not by the campaign going political and its protest potential.
When we mention the volatility of Russian voters, we do not mean there is no ideological voting. On the contrary, the party's ideological relatability is one of the more significant factors for many voters. At the same time, consistent voters become inconsistent for two reasons: if their preferred party does not pay attention to the issues that concern the voter and if the party disappears from electoral space. "If a party for one reason or another disappears from the space, this will affect those voters, whose own position is identical with, or not far removed from the position of the disappearing party, and the result will be that the average probability of vote transfers will increase, ceteris paribus" [27:16]. In other words, when the preferred party is not on the ballot, electoral volatility naturally tends to increase.
As for the possibility of discouraged turnout for United Russia, it can be explained by Morris P. Fiorina's theory of cross pressures. When strong party loyalty conflicts with the party's stance on a relevant issue (in our case, socio-economic deterioration was the issue), the rule is that an undecided voter is more likely to stay home. The less loyal voters often choose protest voting as an alternative [7: 402].
Finally, the decision to vote and the choice of a specific candidate could be two unrelated processes. Let us venture a guess that by the time the "Smart Voting" list was revealed the voters had already decided on voting opposition. Mass protests triggered by candidate disqualifications and political repressions created more stimuli to go and vote than confidence in the "smart" candidate's victory. The larger the scale of the protests and the wider their coverage is, the more citizens are urged to share the displeasure and demands of the protesters . Weekly unauthorized marches all over the city made many Moscow citizens aware of the protests as well as making them witnesses and even unwitting participants while police brutality only fueled the disaffection and urged the citizens to identify with the protesters.
At the same time, it is important to distinguish Navalny's mobilization capabilities from "Smart Voting" itself, as these are completely different research objects. Navalny's supporters were more active in street protests indeed, but the latter would not have been as successful without both registered and unregistered candidates. Navalny's team was represented by five disqualified candidates only while the prosecutor's office was investigating 15 candidates, whom the authorities considered to be behind the unauthorized rallies. The largest rally held on August 10 on Sakharov Avenue was organized by municipal deputies while famous journalists and musicians with extensive online following called to participate in the rally. Analyzing the online activity of Navalny's supporters reveals a correlation with protests, although researchers do not yet have any proof of it being direct: "We can carefully suppose that online activities of Navalny’s supporters increase probabilities of such protests in those regions where it is already intense" [21: 2].
Protest cannot be stimulated at a leader's whim. A leader can only maintain the protest level for some time, if there are all the necessary economic, psychological and social conditions for it. According to Ted Gurr, the main reason behind political discontent is people's frustration with their government figures, who make them suffer losses, and lost hope is the best incentive for protest and rioting . In other words, bad policy and government's mistakes stimulate protest, not actions of the opposition, however effective they are.
Election itself is a powerful incentive for mobilizing dissatisfaction. Like any emotional reaction, protest can emerge from spontaneous expression of indignation, but "protest [itself] is not really what motivates their choices on Election Day": "protest voting seems to be a direct result of political distrust" [3: 369-376]. Distrust has to emerge before indignation is expressed at the polls. This is why protest should be studied in a wider context.
Protest mobilization is cyclic and fluctuating . More than 4.7 thousand protests of differing scale took place in Russia between 2007 and 2012, averaging 700-800 a year. The number of protests is rising in 2008-2009 (the time of financial crisis) and then gradually dropping and rising again by the end of 2011 [16: 337]. Protests are peaking in 2012-2013, with over 2000 each year. At the time, the opposition candidates are highly successful in the elections. Mikhail Prokhorov scores 8% in the presidential election (20% in Moscow and over 15% in Saint Petersburg), Alexey Navalny scores 27% in the Moscow mayoral election. Yevgeny Roizman, Galina Shirshina, Yevgeny Urlashov and Anatoly Lokot become mayors of Yekaterinburg, Petrozavodsk, Yaroslavl and Novosibirsk respectively while Boris Nemtsov is elected to Yaroslavl Oblast Duma.
The wave of protest dwindled after Russia’s seizure of Crimea and gave way to patriotic resurgence instead. Territorial expansion and Western sanctions boosted the regime's legitimacy and created a "rally round the flag" effect . The number of protests in Russia dropped by 40 pp in such circumstances. The 2014 Moscow election results, along with subsequent legislative and presidential election, reveal a level of rallying around the regime, decreased support for the opposition, a readiness to endure domestic trouble for the sake of the "great nation," but also arouse great expectations from the government.
By summer of 2018, the "Crimean Consensus" had already exhausted its mobilization potential. The retirement age increase, the fall of real income and continued economic recession had brought the domestic political agenda back into public spotlight and shaken the faith in the regime. From this moment on, the frustration and appetite for change started to grow while the social demands often started to turn into political more and more often and reflect off election results. Social frustration fueled the protests: in 2018-2019, their numbers spiked to over 2000 protests a year once more . For the first time in 10 years, gubernatorial elections in three regions took two rounds, and the opposition candidate won each round. United Russia shows the worst results in legislative elections since 2007.
As can be seen from this brief overview, the protest had grown quite strong before the 2019 campaign. The results of the previous election cycle "created positive stimuli for the opposition" that, when combined with socioeconomic issues and mistakes of the government, were "able to ignite the desire for political engagement in Russian citizens by "accidentally" causing democratization" [35: 44]. However, Moscow protests were not an isolated event, but part of the bigger wave of political mobilization. This wave can be compared to the 2011-2012 protests for fair elections and includes anti-landfill protests, campaigns against raising the retirement age and healthcare optimization, blocking Telegram and the arrest of journalist Ivan Golunov. The wave is made of three elements: economic, political and local, the latter relating to urban development and local issues.
At the same time, social and democratic agenda came back to Moscow politics before they did to federal, owing to the renovation program and municipal election, which the opposition treated seriously for the first time ever. As a result of the 2017 campaign, the party of power lost its majority in 25 districts while the elected deputies rose above the municipal level by becoming visible actors of city politics, liaisons between local, social and political agendas. Civic activists distance themselves from politics less and less by appealing to political parties for support and putting forward the claims that the government should be replaced. In its turn, political opposition becomes more reliant on social populism and resolving economic issues.
We disagree that economic protest in the Moscow election was irrelevant [29: 97]. The consequences of the government's economic decisions manifest gradually while frustrations continue to grow and burst during major political mobilizations. Election day is the most opportune time for it. In this sense, electoral behavior of Russian citizens gets back to normal when they reward or punish the government based on their personal assessment of socioeconomic policy and their own wellbeing .
In conclusion, we would like to stress that electoral rise of the opposition is not an accident, but a consequence of change in public attitudes and voter preferences. "Smart Voting" did not lay the foundation for the changed choice and votes overflow to alternative candidates. This government itself brought the situation about by introducing unpopular reforms, and disqualification of candidates only reduced the competition among the opposition and, on the contrary, boosted it among the main alternative and administrative candidates. We recognize the impact the new strategy had on the election results, but insist it was not as significant as is commonly believed, and that it played a controversial role in certain constituencies by helping some strong candidates, yet holding back other. The results in certain constituencies would have been different had "Smart Voting" not been used there, but the overall outcome of election and the structure of Moscow parliament would hardly have changed.
Electoral behavior of Moscow citizens changed significantly: the levels of protest and tactical voting increased compared to previous campaigns. These voting methods are different by nature, motivations and goals, but they typically helped the same candidates candidates win, which led to some analysts discussing the great success of "Smart Voting" and attributing all the extra support received by alternative candidates to it. Unlike these analysts, we mark the distinction between these methods and assume that "Smart Voting" was an example of tactical behavior, which should be assessed separately from protest behavior.
At the same time, we do not believe our approach is infallible, nor do we believe that we have covered all the possible impact the "Smart Voting" strategy had on the election results. We gave insufficient consideration to how much the choice of voters with changed preferences depended on the quality of a candidate's campaign, his or her personal characteristics or other expressive reasons. We did not clock any of tactical or protest voting behavior in a number of constituencies, although circumstantial evidence indicates their presence. For instance, out of the seven constituencies that registered increased support for administrative candidates, the biggest surplus of votes went to the journalist Roman Babayan (+14.62 pp) and Ilya Sviridov (+7.67 pp), the ex-mayoral candidate from A Just Russia – this surplus was more of a sign of protest than loyalty. Our model is built on tactical voting for the opposition candidates, although it is possible that pro-government voters could have also made a tactical choice. We are also aware that many voters had difficulty distinguishing between the administrative and alternative candidate, which influences their choice as well. Although our model has some other limitations, we do hope that it provides a more detailed explanation of the specifics and principles of the 2019 Moscow election campaign than other existing models do.
Received 05.05.2020, revision received 12.05.2020.