Electoral Cleavage Structure: 2021 State Duma Election

Korgunyuk Yu.G.

Abstract

The paper presents an analysis of electoral cleavages at the federal and regional levels following the 2021 State Duma election. There is a need to shift the focus of the cleavage theory to studying the situational dynamics of mass political consciousness. The author concludes that the factors of interregional sociodemographic differentiation affect voter attitudes in a complex manner instead of each factor having an isolated effect. The author therefore proposes to abandon the approach where these factors are viewed as a source of political confrontations, stressing that it is political actors who initiate the confrontations in an attempt to expand their social base. In the setting of electoral authoritarianism, these political actors are overwhelmingly represented by government elites, who have initiated such confrontations as "authoritarianism–democracy", "hawks–doves", and "Soviet traditionalists–progressives". It is the elites' proactive approach, not the social base of electoral cleavages, that makes their political content more fluid and changeable. The constants of the 2016 and 2021 electoral cleavage structures include the dominance of authoritarian-democratic confrontation, whose social base is more related to the level of urbanization, but at the same time to demographic characteristics and the population's level of economic independence. Another constant—a consistent second—is the socioeconomic confrontations between government and opposition as well as the free market supporters (market liberals) and social protectionists. The former is gradually growing its momentum, while the latter is losing (because it is pushed out by confrontations on foreign policy and worldview issues: "hawks" vs."doves", conservatives vs. liberals, Soviet traditionalists vs. progressives). The novelty of the 2021 elections is the expanded range of political confrontations that fill electoral cleavages with content, including those associated with the emergence of new parties.

---

At present, the cleavage theory is going through some tough times. Compared to 1960–1980, it has been getting much less attention. There are several reasons for this.

One of them is that the theory aimed to study political processes at the macro level. It operated for decades and even centuries, painting in broad strokes a picture of how party systems developed.

The theory's original version [27] attempted to explain the evolution of party systems in Western Europe (in its northern part, to be exact) as one based on fundamental conflicts of social interests that persisted over time. The conflicts were said to arise between 1) center and periphery (i.e., capital and regional elites; 2) state and the church (essentially, between religious and secular voters); 3) land owners and industrial owners (at a later stage, this conflict came to be conflated with that of urban vs. rural areas, with little to no reasoning behind it); and 4) owners and workers.

One of the key points of the cleavage concept developed by Lipset and Rokkan is the hypothesis that after universal suffrage was introduced in Europe, the established party systems "froze" for several decades — until the 1960s. Cleavage theory accordingly argued that the relative stability of Western European party systems was ensured by the dominant conflict — that of owners vs. workers.

The irony is that by that time there were already signs of "unfreezing" in European party systems. In the following decades, the process unfolded in full force. In today's West, it is not easy to find a country where the party-politics balance of power has not undergone significant changes, accompanied by former leaders losing momentum and new leaders gaining one. One of the few exceptions is the United States of America with its unsinkable two-party system. Yet it is not the "frozen" cleavage structure that keeps the system afloat there either. Kenneth Janda's work in particular shows how through the course of their existence, the political facade and social base of the country's two leading parties changed not once, but several times, and dramatically at that [11; 12].

Cleavages were viewed as the cornerstone of a party system's stability, yet when this view grew muddled, party researchers started losing interest in the theory. On the other hand, researchers of mass political consciousness and voter values retained their interest [2; 9; 25; 29; 30]. For these researchers, parties were only agents involved in shaping mass political consciousness, which, on the one hand, is conditioned by voters' social status and, on the other hand, operates with clichés that active political players push in the media.

In this framework, cleavages are merely political confrontations with social content. Parties are engaged in generating them for their own benefit: a political agenda can only come through if it touches a sore spot in a voter's mind. Requiring the set of such irritants to remain unchanged is by and large excessive.

Therefore, with cleavages defined as "long-term structural conflicts that lead to the emergence of opposing positions" [26: 51], it is quite acceptable to disregard the "long-term" condition. Cleavages can be of any length, as long as they are have a political interpretation and social content. After all, scholars of social and political processes are quite accustomed to either being blind men touching an elephant or inhabiting Plato's cave and trying to guess what the shadows on its wall represent. It is possible that the cleavages the scholars record are reflections of larger rifts, but only time can tell what the latter are.

In other words, in the new century, cleavage researchers have been faced with a choice: either to abandon this theory altogether, or to move to the micro-level, which in this case is not the level of the individual, as is the case in social studies, but the level of an individual election. There is no paradox to this, because each subject of research has its micro- and macro-levels, and what is macro-level for one subject is a micro-level for another (for more see [18]).

In this sense, cleavages may reflect not only long-term trends in the development of the social structure of society, but also quite opportunistic changes in the agenda, caused by unforeseen factors: natural disasters, epidemics, geopolitical conflicts, etc.

Naturally, one could coin a new term for such phenomena. The only question is why do that when a befitting term already exists, even if it is collecting dust at present? After all, no phenomena and processes can exist at the macro-level alone; each has its micro-level that looks very different up close than it does from afar. Using the term "cleavage" to describe phenomena more volatile and unstable than "long-term structural conflicts" is thus appropriate if only out of respect for Occam's razor, which calls not to multiply entities beyond necessity.

The fact that cleavages are not as stable and persistent as Lipset and Rokkan's theory implied became clear fairly quickly. The fact is that the foundational paper published in 1967 barely mentioned the political aspect of cleavages. For example, it is clear that the conflict between workers and owners involved a confrontation between socialist (communist) and "bourgeois" parties. But it only looked this way from the "left" side of the political spectrum, not from the opposite flank. All the more so because there was no united opposite flank, but a combination of ideologically diverse parties, from monarchists to liberals. From the point of view of monarchists, socialists and communists were a radical republican faction; from the point of view of conservatives, they were radical liberals; from the point of view of liberals, they were supporters of tyranny and demagogues who used pseudo-democratic slogans. In short, cleavage structure was much more complex, clearly going beyond the confrontation between socialists and non-socialists. One can quite understand Lipset and Rokkan, who chose not to go into details, but to abstract from them.

However, for those who tried to apply Lipset and Rokkan's pattern to countries outside the theory's initial range, it was clear that the political component of cleavages clearly required further study. These researchers quickly shifted their focus from the social component of cleavages to the "ideological" one. The researchers in question first of all include the participants of the Manifesto project, who studied issues from electoral platforms and programs of political parties. Their research laid the foundation of the issue dimensions theory [4].

The project's analysis produced rather mixed results, revealing more dimensions than was initially anticipated. In many respects this was due to methodological shortcomings (for details see [16]), but it did not stop at that: the political content of the confrontations within the party systems did not fit into the Procrustean bed of the initial premises and required a broader descriptive approach.

The experience of post-Communist countries in Central and Eastern Europe attracted even greater interest to the political component of cleavages. In this case, the set of cleavages varied from country to country. Moreover, the cleavages themselves were not particularly stable and evolved rapidly. This fluidity brought one particularly excellent work into existence. A study of post-communist party systems by Herbert Kitschelt, Zdenka Mansfeldova, Radoslav Markowski and Gábor Tóka suggested terms like "divides" and "dimensions" instead of "cleavages," arguing that the former, were less stable than the latter [13: 63]. But this seems to be the cost of an approach where the main point of cleavage theory is to explain the stability of party systems; if the stability criterion is abandoned, the suggested terms will easily overlap with cleavages.

One way or another, the book focused more on the political content of the "divides" and "dimensions" than on their social base, which is why its authors relied on expert reviews and public opinion polls instead of election results as a source of material. As a result, post-communist Czech Republic, Poland, Hungary, and Bulgaria revealed such dimensions as "social protectionists vs. market liberals," "libertarian-secular versus authoritarian-religious" [13: 226], "political libertarian–authoritarian division" [13: 235], "socio-cultural libertarian versus authoritarian" [13: 259] and some others defined by local context.

The study demonstrated that the political content of "divides" (dimensions) changed not only from country to country, but also from election to election. This is largely due to the fact that its authors were much closer to the micro-level than Lipset, Rokkan and their followers, who saw cleavages as the source of institutionalization of party systems. In their cleavage studies, Kitchelt and his colleagues used more subtle mathematical methods and were initially prepared to detect patterns that deviated significantly from "classical" benchmarks.

As for the social foundations of the identified cleavages, they did not find the same rich diversity, concluding that "young, male, well-educated, urban voters in the private sector and in high-skill jobs should be most inclined to support market-liberal policies" [13: 296].

A similar pattern emerged in post-Soviet Russia. The question arises, however, whether cleavage theory can be applied to a country whose political regime is becoming increasingly authoritarian from election to election. Can it be applied to such regime types in general? The answer is: why not, if votes are actually counted in a more or less substantial number of territories. While administrative promotion of candidates endorsed by the "party of power" and even blatant ballot stuffing does distort the electoral space and cleavage structures, it does not eliminate them completely.

To begin with, administrative intervention itself brings forth the confrontation between the government and the public — one with an authoritarian-democratic content. In other words, it is a confrontation between those who are able to resist the administrative pressure and those who are not. This rift can be quite reliably captured by mathematical methods in particular [1; 14; 18].

We should also point out that electoral autocracies rarely resort to all-out electoral fraud. They slip extra votes to their candidates and parties, but they do not tamper with the results of the rest of the participants — because that requires extra work. As for cases of all-out fraud, it is easy enough to reveal them by doing a mathematical analysis of vote returns; they end up having no intelligible interpretation.

Typical ballot-box stuffing distorts the balance of forces between the election participants, but does not actually change it. This means that when applied to vote returns, the same mathematical analysis first, may reveal 1) whether there was full-scale fraud involved, and 2) cleavage structure similar to the cultural layer in archaeological excavations: covered with soil, but only partially damaged.

When comparing cleavage structure in Central and Eastern Europe and in post-Soviet Russia, it is quite obvious that in terms of its social base, the structure is not diverse and fluid either: it is primarily influenced by the urbanization level, the demographic (including ethnic) composition of the territories and the level of economic (including entrepreneurial) activity of the population [1; 15; 23].

In political terms, on the other hand, the transformation of cleavages is impressive in both diversity and speed. While studying the inter-party debate during the State Duma elections in 1993-2021, the author used factor analysis to identify three main dimensions in Russian political space: 1) socioeconomic; 2) authoritarian-democratic; and 3) so-called systemic [14; 15].

The first and the second are more or less clear. The former is referred to as "right-left" by Ian Budge and "socioeconomic protectionists vs. market liberals" by Herbert Kitchelt et al.

The authoritarian-democratic dimension is not very relevant in the West today, but it is quite widespread in post-Soviet countries and Latin America [1; 32: 32; 34]. As of recent, it has once again manifested itself in Central and Eastern Europe [31]. In general, the authoritarian-democratic dimension manifests itself in those countries where there are either growing trends toward electoral authoritarianism (Russia and post-Soviet countries), or where the memory of a recent dictatorship is still alive (Chile) [34].

The third, "systemic" dimension generally corresponds to the sociocultural dimension of Western democracies [2; 25; 29; 30]. It may be considered a battleground between advocates of open/closed societal systems. This openness/closedness is both temporal and spatial: in the first case the opponents are either future- or past-oriented, while in the second they aim at either integration of supranational organizations or, on the contrary, at isolationism.

The temporal version can be observed in the confrontation between materialists and post-materialists [10], new and old politics [5]. The spatial version manifests itself in the confrontation between demarcationists and integrationists (i.e. those who have benefited from globalization and those who have not) [24], universalists and particularists [6], cosmopolitans and communitarians [33]. There are also combined versions that contrast libertarian universalists with traditionalist communitarians [3] or greens/alternativists/libertarians with traditionalists/authoritarians/nationalists - GAL/TAN [7; 8].

If we limit ourselves strictly to inter-party debate, then between 1993 and 2011 the hierarchy of political dimensions in Russia remained relatively stable: the socioeconomic dimension almost always came first, the systemic dimension second, and the authoritarian-democratic dimension third. In 2012, the authoritarian-democratic dimension took the lead, pushing the socioeconomic and systemic dimensions to second and third place, respectively (a direct consequence of the "white-ribbon revolution" of 2011–2012 and the subsequent tightening of the repressive legislation [14; 16]). In 2014, the Ukrainian crisis brought to the forefront a systemic dimension: the "hawks," or imperialists (the vast majority of Russian parties), vs. the "doves," or anti-imperialists (exclusively liberal). The authoritarian-democratic and socioeconomic dimensions moved to second and third place, respectively [16].

But this applies first and foremost to political space, which is defined by inter-party debate and reflects the consciousness of the elites, who have a sufficiently coherent, though subjective, view of this space. The electoral space, which lives by the laws of mass consciousness, is a different matter altogether. The mass consciousness is quite fragmented, since most voters are unable to perceive the political landscape as a whole, perceiving it through their own social problems instead; hence the incomplete congruence of political and electoral spaces.

In the political space, the socioeconomic dimension (coupled with elements of the worldview confrontation between Soviet traditionalists and pro-Western progressives) dominated until 2011. In the electoral space, however, it was overshadowed by the authoritarian-democratic one as early as the beginning of the 2000s. The main reasons for this were the Kremlin centralizing administrative resources (in 1990, they were mainly concentrated in the hands of regional elites), increasing administrative pressure on voters, increasing electoral manipulation and direct fraud. The situation did not change after 2014, when the confrontation between "hawks" and "doves" (systemic dimension) came to the forefront of the political space. In the 2016 State Duma election, the dominant electoral cleavage retained its authoritarian-democratic content, while the imperialist-anti-imperialist dimension ("hawks and doves") continued to define the second most important electoral cleavage together with the socioeconomic dimension [16].

One way or another, we may state that the relative invariability of the cleavages' social base coexists with the fluidity of their political component, which is explained by the desire of political (primarily power) elites to impose their agenda on the whole country. This can be observed in post-Soviet Russia, but it is very likely to be the same in other countries, especially given the tumultuous political change of recent decades.

The years 2000 and 2014 may be considered the critical points in the transformation of post-Soviet Russia's political space. The first point was marked by centralization of administrative resources, which brought the authoritarian-democratic confrontation into the mass political consciousness. At the second point, the power elite gave the political space a powerful injection of foreign policy agenda; this did not affect the dominant electoral cleavage, which retained its authoritarian-democratic nature, but it did affect the second EC, diluting its socioeconomic content with elements of the confrontation between "hawks" and "doves.

In other words, we see the political elite constantly imposing the current agenda. At the same time, the mass consciousness either succumbs to this pressure or demonstrates inertness (or even resistance). As a result, cleavage structure appears to be relatively stable socially and fluidly unstable politically. It is possible to determine the trends in its transformation only by carefully and consistently studying the changes in the relationship between the political and social component (including how their dynamics changes from election to election).

Such a study can take many decades, but Russia's regional diversity is particularly helpful for researchers in this case. Russia's federal subjects include regions that may seem quite modern to a European eye (Moscow, St. Petersburg), and quite "patriarchal" (mostly the republics in the North Caucasus). It is possible to describe the differences between these categories using an unconventional counterpart to a dynamic model, where "patriarchal" and "modern" regions correspond to different cleavage structures. These very differences provide ample material for analysis.

This paper uses cleavage theory to study the results of the 2021 State Duma election at the federal and regional levels. The methods employed by the author have been described in previous papers on the same subject [18; 17; 20]. Comparing present study results with the 2016 results, the author draws conclusions about the cleavage structure trends.

Research methodology

In many respects, the research methodology replicates the one used in previous studies.

Factor analysis (principal component analysis) was chosen as the main method for identifying cleavages both in the electoral and political space. Its R-mode was chosen specifically as it is geared toward grouping of variables instead of objects (as in the case with the Q-mode). Factor analysis is relevant for identifying cleavages because it helps to reduce the number of factors by focusing on those that imply a higher dispersion, and therefore, in our case, a higher polarization.

This is an important point for our study because of its subject: voting for party lists in various territorial units. Factor analysis of the number (shares) of votes for different parties reveals the main patterns of party voting in said units by showing whether party results change from territory to territory in a synchronized manner and or in counter-phase. To be fair, correlation analysis may detect the most striking of these patterns. However, it captures only the most obvious connections, which is why factor analysis fits best as it detects second- and third-order connections.

Factor analysis also helps to determine the hierarchy of these patterns. For example, it helps to determine that the first pattern is the confrontation between the "party of power" + its allies and the most opposition parties; the second pattern is the confrontation between communist and liberal opposition blocs, etc.

For this reason, our study begins with factors of differences in voting for parties in different territorial units. The factors were calculated by factor analysis of the number (shares) of votes received by parties in federal subjects and territorial constituencies (depending on the level of research: federal or regional). Parties acted as variables, whereas territorial units acted as cases.

If these factors were found to have political and social content, they were regarded as electoral cleavages.

The political content was established by comparing the factor loadings of parties in the electoral and political spaces using correlation and regression analyses. Factor loadings in the political space were calculated based on the parties' positions on the most polarizing issues. These issues, as well as party positions, are identified through monitoring that the author has been conducting since 2015; the relevant documents and materials can be found in the "PartyArchive" database, http://www.partinform.ru/pa98. Party positions were evaluated on a scale from –5 to +5 and were also subjected to factor analysis, where the parties were taken as variables and the issues were taken as cases (for more details on the methodology, see [15; 23; 16; 18]).

This is another reason why factor analysis was chosen, as it allows us to reduce the numerous party confrontations on a wide range of issues to some more compact patterns, mainly those characterized by the greatest polarization.

Factor analysis was performed both on the whole set of issues of inter-party debate and in separate issue domains: 1) domestic political; 2) socioeconomic; 3) foreign policy; and 4) worldview. The last two domains (foreign policy and worldview) are considered separately and together in different models in order to increase the number of possible configurations. It is impossible to predict at what points a fragmented mass consciousness will be congruent with a relatively coherent elite one, so it is better to have multiple options.

Three models were used for this purpose. Model 1 contains the main political dimensions (they largely correspond to "issue dimensions" described by Ian Budge and his colleagues in the "Manifesto" project) for the whole set of inter-party debate issues. Model 2 contains confrontations in three issue domains (a term adopted from Ian Budge et al.): domestic policy, socioeconomic and "systemic" (foreign policy + worldview). Model three contains confrontations in four subject areas (all of the aforementioned considered separately).

The identified political dimensions and sub-dimensions (confrontations in issue domains) were interpreted depending on which parties were the main antagonists in each of them (indicated by the most polarized factor loadings) and which issues caused the greatest polarization (and therefore had maximum modulo factor scores above one, as a rule). Multiple regression was also used to interpret the main political dimensions, where the dependent variable is the party factor loadings on the entire set of issues, and the independent variable is the party factor loadings on the confrontations in specific issue domains. This helps to determine the "weight" of specific sub-dimensions within the main political dimensions and to understand which "shades" dominate their content. In some approximation, this can also be found out by correlation analysis, but multiple regression allows us to take into account the possible cumulative effect of several correlations overlapping.

To find out whether the factors of territorial differences in party voting have any political content, the factor loadings of parties in the electoral space were compared with their factor loadings within political dimensions and sub-dimensions. For this purpose we used multiple regression, where factor loadings in the electoral space were taken as dependent variables, and factor loadings in political dimensions and sub-dimensions were taken as independent variables. It is possible to detect partial relationships between factor loadings of parties in the electoral and political spaces by correlation analysis, but multiple regression, as in the previous case, allows us to take into account the cumulative effect of several correlations overlapping.

This study's regression analysis was conducted in three models as well. The factor loadings of parties acted as independent variables: in the main political dimensions in Model 1, in three issue domains (domestic policy, socioeconomic and systemic) in Model 2, in four subject issue domains (domestic policy, socioeconomic, foreign policy, worldview) in Model 3.

If any of the models gives coefficients where statistical significance is below 0.1, then it is assumed that there is an electoral cleavage with political content. In fact, a threshold of 0.05 is used as the more exact main benchmark, but the statistical significance requirements are less strict if there are no results following the "something is better than nothing" principle. Given the complex nature of socio-political phenomena, one should not become excessively rigorous when studying them as it is better to record a weakened connection than to record none at all. After all, statistical methods in the social sciences have no evidentiary force and are primarily a heuristic tool.

Multiple regression was also used to determine whether the identified electoral cleavages have a social base. This time, however, the factor scores of territorial units in electoral cleavages were taken as the dependent variable, and the demographic and socioeconomic indicators of the same territorial units were taken as the independent variables. The latter indicators were subjected to factor analysis beforehand (the main factors of socio-demographic differentiation typically included the quality of life, which almost coincided with urbanization level; demographic characteristics; the level of economic independence of the population) [15; 18; 17]. If the regression model revealed correlations with statistical significance below 0.1, it was assumed that the electoral cleavages had a social base.

Furthermore, electoral cleavage structure is described using measurement tools proposed in the author's previous works [18; 17]:

– EC maximum range coefficient formula: \(Mc = \sum_p^n |FLp|\), where \(p\) is the share of votes received by each party under the proportional representation system; \(n\) is the number of participating parties; \(|FLp|\) is the factor loading modulo for each party on the given cleavage;

– EC effective range coefficient formula: \(Ec = 2Mc_{min}\), where \(Mc_{min}\) is the range of the weaker side of the cleavage;

– EC politicization coefficient formula: \(Pc = Ec R^2\), where \(R^2\) is the multiple regression coefficient (coefficient of determination) of the relationship between the EC and a set of political dimensions and domain cleavages;

– EC socialization coefficient formula: \(Sc = Ec R^2\), where \(R^2\) is a coefficient of determination that reflects the connection of each EC with a set of social stratification factors typical for the given region.

The interrelation of these indicators demonstrates the level of competition in elections, the degree to which the electorate understands the landscape of political confrontation and how vote returns are conditioned by the social status of voters.

As a result, the close values of maximum and effective range coefficients indicate a high level of competition, while strongly discrepant values indicate the privileged position of one of the participants. A high politicization coefficient value indicates that voters have a clear understanding of the alignment of forces (as in voters understand which parties oppose each other on issues relevant for them), while a low value indicates that they take little interest in it. A high socialization coefficient means that a voter's choice is largely determined by his or her social status, while a low one means that this fact does not have any significant impact.

That said, the focus is on complete electoral cleavages — those that have both political and social content. After all, it is through them that we can understand how the social status of voters affects their political choices and which factor of social differentiation "generates" a certain type of political confrontation.

In conclusion, the results of the 2021 State Duma elections are compared with the same results from 2016.

The Configuration of Political Space in the 2021 Election: Political Dimensions and Domain Cleavages

The inter-party debate in the 2021 Duma election revealed 186 issues that caused a more or less serious polarization of party positions. All these issues were selected through daily monitoring of public speeches (interviews, debates, etc.) of party figures, as well as the content of openly published party documents (programs, platforms, statements, appeals, etc.). We only considered the issues that were discussed by a more or less significant number of participating parties. The declared positions always included polarized ones (with values of +5 and -5).

Factor analysis of the positions of the 14 parties admitted to the elections revealed four political dimensions (Table 1). This was a surprising development since in previous Duma campaigns there were always three.

It is clear that the main content of the first dimension is the confrontation between United Russia and all other participants, but primarily the CPRF, Communists of Russia, LDPR, A Just Russia – For Truth, and Rodina; the second dimension contains the confrontation between United Russia and liberals (primarily Yabloko); the fourth dimension contains the particular position of Green Alternative. However, let us not be too hasty with interpretation, and instead look at the analysis results for each issue domain.

The "Domestic policy" issue domain included 51 issues of inter-party debate out of the above-mentioned 186. The selection criterion for this domain was the topics related to the political structure of the state, the conduct of elections, the functioning of government and other political institutions, etc.

Factor analysis identified three confrontations (sub-dimensions):

1) government vs. opposition (United Russia vs. the rest, primarily CPRF, Communists of Russia, Yabloko, Party of Growth, New People); the most polarizing issues: government use of administrative resource in elections; expanding the powers of the representative branch; popular election of local self-government head officials; abolishing municipal filter; cutting down proportional representation; Internet restrictions; multi-day voting; restrictions on holding rallies (Table 2);

2) loyalists vs. opposition (Rodina, REP The Greens, Civic Platform, etc. vs. Russian Party for Social Justice, Yabloko and CPRF); polarizing issues: assessing Russia's political regime as authoritarian, views on Alexey Navalny and his projects, views on Vladimir Putin and his amendments to the Constitution;

3) Green Alternative's particular position on a number of issues, mostly related to coronavirus response measures.

The socioeconomic domain included 54 issues, and factor analysis revealed four sub-dimensions:

1) government vs. opposition (United Russia vs. the rest, primarily CPRF, Communists of Russia, Rodina, A Just Russia – For Truth, and LDPR); polarizing issues: changing back the retirement age, abolishing capital improvement payments, views on "optimization" of healthcare and education, on waste management reform, on the Central Bank and government policies, etc.;

2) liberals (New People, Party of Growth, Yabloko) vs. Green Alternative: waste management reform, distribution of natural resource revenue (universal basic income) among citizens, redistribution of taxes in favor of regions and local governments, and some others (GA simply had no position on most of these);

3) liberals and environmentalists vs. social protectionists (CPRF, Communists of Russia, A Just Russia – For Truth, Russian Party of Pensioners): waste management reform; distribution of natural resource revenue; advantages of private property over state property, revision of 1990s privatization;

4) new parties (NP and GA) vs. loyalists (United Russia, LDPR and Civic Platform): natural resource revenue, reducing military and defense spending, price and tariff freezing, status of self-employed workers, reindustrialization.

Systemic domain included 90 issues related to foreign policy and worldview subject areas. In this case, factor analysis revealed four sub-dimensions:

1) "hawks" vs. "doves" (communists and "patriots" vs. Yabloko, with the New People and Green Alternative neutral): Crimea, Ukraine, support for Putin's foreign policy, priority of Russian law over international law; Malaysian Boeing, ban on LGBT propaganda, etc.;

2) Soviet traditionalists (CPRF, CoR) vs. progressives (Party of Growth, Yabloko, REP The Greens): changing Soviet geographic names, rejecting capitalism, the October Revolution, Russia's "own way" of development, bringing back the "ethnicity" clause in passports, etc.

3) New People vs. Green Alternative, Russian Party of Social Justice, and Civic Platform: COVID19-related restrictions, Russia's "own way" of development, the need to take a harder line in foreign policy;

4) New parties (RPFJ, GA, NP) vs. United Russia: laws on foreign agents and undesirable organizations, state control of educational activities.

Systemic issue domain was divided into foreign policy (46 issues) and worldview (42 issues) domains. In the first area, factor analysis revealed four sub-divisions:

1) "hawks" vs. "doves" (roughly the same confrontation and the same polarizing issues as in the corresponding sub-dimension of the systemic issues);

2) authoritarians vs. democrats (CPRF, United Russia, Rodina vs. environmentalists, RPFJ and liberals): the law on undesirable organizations, the 2020 events in Belarus, etc.

3) New People vs. RPFJ and A Just Russia – For Truth: control over foreign social networks and IT resources, the need for a tougher foreign policy, Russia's turn to the east, etc.

The worldview domain revealed four sub-dimensions:

1) liberals (Yabloko) vs. conservatives (Communists, Rodina, A Just Russia – For Truth): protection of traditional values, persecution for LGBT propaganda, abolishing Unified State Exam, patriotic education of youth, etc.;

2) Soviet traditionalists (CPRF and Communists of Russia) vs. progressives (Party of Growth, REP The Greens, New People): attitude towards Russia Day (June 12), advantage of private property over state property, Russia's "own way" of development, rejection of capitalism, the October Revolution;

3) New People vs. Rodina, LDPR, Green Alternative, United Russia: COVID19-related restrictions, Russia's "own way" of development, state control of educational activities, attitude to Yeltsin;

4) United Russia vs. RPFJ and Civic Platform: divergence on minor points.

As we can see, quite a large part of the sub-dimensionality stems from the particular position certain new parties have on certain issues. Some parties (New People, Green Alternative) simply had no opinion on issues important to the main participants in the election. On the other hand, the Russian Party for Social Justice held some rather exotic opinions: it strongly opposed the government on domestic policy issues, took a hawkish position in foreign policy, came close to the communists on social and economic issues, and was not particularly consistent in its worldview.

By applying multiple regression where the factor loadings of parties in the main political dimensions acted as the dependent variable, and the same loadings on the sub-dimensions (confrontations in separate subject areas) acted as independent variables we discovered that in both models (a reminder that in M1 the foreign policy and worldview issue domains are combined into a single systemic one and considered separately in M2) the content of the first two main political dimensions largely coincided: "hawks" vs. "doves" and government vs. opposition were main components (Table 3). The only difference is that in the first dimension, the confrontation between government and opposition extended to both the domestic policy and socioeconomic domains, while in the second it was limited to the domestic policy one. The confrontation between Soviet traditionalists and progressives was the main component of the third political dimension. Market liberals vs. social protectionists in the socioeconomic domain was the main component in the fourth one (in the second model, however, this confrontation gave way to Green Alternative's particular position on domestic policy issues).

At the same time, if we remove the parties with the least votes under proportional representation from the analysis (Civic Platform and Party of Growth), the political space becomes three-dimensional yet again as only three main political dimensions remain.

That said, the first two remained complex: the first one combined elements of confrontation between "hawks" and "doves" as well as between government and opposition in the domestic policy and socioeconomic domains (although in the second regression model—with four domains as predictors — the main elements were government vs. opposition in the socioeconomic domain, and liberals vs. conservatives in the worldview domain). In the second dimension, the government vs. opposition confrontation in the socioeconomic domain fell out of the set (Table 4). As for the third political dimension, in both models it was primarily shaped by the sub-dimensions associated with the new political parties (New People and Green Alternative before all else).

At any rate, Russia's political space changed significantly over the five years. In 2016, there were only three main political dimensions: the main content of the first was the confrontation between "hawks" and "doves" (echoes of the so-called "Crimean spring"), the second combined the confrontation between government and opposition, as well as conservatives and liberals in domestic policy, socioeconomic and systemic domains, and the third consisted in the confrontation between market liberals and social protectionists in the socioeconomic domain, and liberals and loyalists in domestic policy domain [19].

In 2021, the first political dimension seemed to have split into two: in both of them, the opposition between "hawks" and "doves" remained a leading element, but it was no longer dominant, merging with the confrontation government and opposition. In the first case it merged in the domestic policy and socioeconomic domains simultaneously (hence the opposition between United Russia and "socialist patriots"), in the second case only in the domestic policy domain (hence the opposition between the "party of power" and the liberal Yabloko party). In the third political dimension, the opposition of Soviet traditionalists and progressives came to the fore; in the fourth dimension, the opposition of market liberals and social protectionists remained important, but it faced serious competition from the new parties (New People and Green Alternative).

Political interpretation and socioeconomic content of electoral cleavages in the 2021 election: federal level

Factor analysis of vote returns for all 14 parties in all 85 federal subjects revealed three factors of territorial differences (Table 5). The first consisted in confronting United Russia with all the other election participants, but above all the Russian Party of Pensioners, New People, Green Alternative, Communists of Russia, and LDPR. The second consisted in the confrontation between the "old" market liberals (Yabloko and Party of Growth) and the communists (CPRF and Communists of Russia). The third consisted in the competition between Civic Platform and Rodina, A Just Russia – For Truth and LDPR. Given the spoiler nature of Civic Platform, odds-wise this factor may be considered as junk.

To clarify the political interpretation of these factors, three regression models were built. In all three, the dependent variable was the factor loadings of the parties in electoral spaces, while the independent variables were their loadings in the main political dimensions (Model 1) and the domain cleavages (sub-dimensions): domestic policy, socioeconomic and systemic (Model 2) and domestic policy, socioeconomic, foreign policy and worldview domains (Model 3).

For the first factor of territorial differences, the relationship with the first two political dimensions had the highest determination coefficient (Table 6). As we already know, both of them combined the confrontation of "hawks" vs. "doves" and government vs. opposition. In the other two models, there was also a connection with the confrontation between government and opposition in the domestic policy domain, but its determination coefficient was lower, which is why we used the Model 1 coefficient to calculate the cleavage's politicization coefficient.

As for the second territorial differences factor, Model 1 revealed no significant connections, Model 2 revealed a connection with the confrontation of "hawks" and "doves", and Model 3 revealed a connection with the confrontation liberals and conservatives in the worldview domain (Table 6). That said, the latter yielded a larger determination coefficient, which was used to calculate the politicization coefficient.

The third candidate for electoral cleavage status showed no connection to any of the political dimensions or sub-dimensions, which confirmed the assumption of its junk nature.

Curiously, excluding parties with the minimum number of votes from the analysis increased in the determination coefficients, as well as changed the interpretation of electoral cleavages.

For one thing, discarding Civic Platform and Party of Growth immediately reduced the number of territorial difference factors to two (Table 7), which is not surprising, since the third factor was associated primarily with the high modulo factor loading of Civic Platform. After the number of participants was reduced to eight (those who gained over 1% of the votes), the number of electoral cleavages remained the same (Table 8), and their nature did not change either: the first was the confrontation of United Russia with all the other participants, and the second was the confrontation of liberals (Yabloko) and communists (CPRF and Communists of Russia).

For 12 parties, the interpretation remained essentially the same: the combination of the first and second political dimensions had the highest coefficient of determination, and here this coefficient increased slightly, just as it did for the government vs. opposition antithesis in the domestic policy area (Table 9).

As for the second electoral cleavage, the determination coefficient not only increased, but also changed its interpretation. In the model with three sub-dimensions, the confrontation between "hawks" and "doves" was replaced by the confrontation between market liberals and social protectionists in the socioeconomic domain (Table 14). In the model with four sub-dimensions, the confrontation between liberals and conservatives in the worldview domain was supplemented with a factor connected with New People's particular position on foreign policy issues (which consisted, however, in the party trying to avoid these issues). As a result, the determination coefficient did not simply increase, but literally skyrocketed.

Even more significant were the changes for the eight parties that gained over 1% of votes. In all three models, the determination coefficient for the first electoral cleavage spiked, especially for the four-sub-dimension model. At the same time, its political content slightly changed. If for the model with the main political dimensions it was still conditioned by ECs 1 and 2, then in the models with three and four sub-dimensions its main component was the confrontation between government and opposition in the socioeconomic domain, instead of the domestic policy one. It is just that in the first case, it was laced with the confrontation between liberals and conservatives in the systemic domain, and in the second with the special (essentially non-existent) position of New People on foreign policy issues (Table 10).

Determination coefficient for the second electoral cleavage did not simply decrease: its statistical significance in models 2 and 3 went beyond 0.05 (although it failed to reach 0.1). The political content of this factor also changed: there was a confrontation between "hawks" and "doves" in the 3-sub-dimensions model, and Soviet traditionalists vs. progressives on ideological issues in the 4-sub-dimensions model.

The increase in the determination coefficient (and consequently, the politicization coefficient) for the first electoral cleavage and the decrease for the second indicates that the first EC was associated with the confrontation of the major parties, while the second reflected the effect minor parties had on the mass political consciousness; many of these parties failed to get even 1% of the votes.

It should be noted that all the above-mentioned dimensions and sub-dimensions were relevant to the political content of the identified electoral cleavages: simply put, different electorate groups were more interested in different parts of political agenda.

On the whole, the first two political dimensions played a major role in the political content of the first electoral cleavage, as well as the confrontation between government and opposition in the domestic politics domain. However, there was a marked presence of the confrontation between government and opposition in the socioeconomic domain, as well as of the confrontation between liberals and conservatives in the systemic domain, and even New People's particular position on foreign policy.

The second electoral cleavage was affected by the confrontations between "hawks" and "doves," liberals and conservatives, and Soviet traditionalists and progressives in the worldview domain, between market liberals and social protectionists, and the very same particular position of New People on foreign policy.

Let us now turn to the social base of the identified factors of territorial differences in party voting. To determine the social base, the factor scores of federal subjects in the electoral space were compared with their factor scores, but it was done based on factor analysis results for demographic and socioeconomic indicators of these regions (data for the beginning of 2021 was sourced from the Roskomstat website, https://rosstat.gov.ru/). In total, factor analysis revealed five factors of inter-regional socio-demographic differentiation, which were interpreted as: 1) demographic characteristics; 2) urbanization level; 3) social wellbeing of the population; 4) economic independence of the population; 5) government aid to the population (Table 11).

Regression analysis of the links between electoral cleavages and factors of inter-regional socio-demographic differentiation revealed predictors for the first two ECs and did not reveal them for the third, which once again confirmed its junk nature.

The first electoral cleavage was associated primarily with the level of urbanization, then with the level of social wellbeing, and least of all with demographic characteristics for the 14- and 12-party models, and only with urbanization and demographic characteristics for the 8-party model (Table 12).

For the second electoral cleavage, the strongest predictor was the level of social wellbeing for the 14-party model and the level of economic independence of the population for the 12- and 8-party models (Table 12). In fact, it was the only predictor for the 8-party model, while in the 14- and 12-party models there was also the "Demographic characteristics" factor present, placing last in both cases.

As we can see, the first electoral cleavage was shaped primarily by the level of urbanization (coupled with the level of social wellbeing), and the second by the level of economic independence of the population (also coupled with the level of social wellbeing). Demographic characteristics indeed played a role in both cases, but we will focus mostly on the leading factors.

With a certain share of convention, we may thus argue that the confrontation between government and opposition on domestic political and socioeconomic issues is related primarily to the level of urbanization, while the confrontation between liberals and conservatives (market liberals and social protectionists), Soviet traditionalists and progressives, "hawks" and "doves" are related both to the economic activity of the population and its social wellbeing.

Table 13 shows the coefficients of maximum and effective range as well as politicization and socialization of electoral cleavages based on three models: for 14, 12, and 8 parties.

For all models, the maximum and effective range, as well as the socialization coefficient of the first electoral cleavage fluctuate around the same values: the maximum range about 81, the effective range about 65.5, and the socialization coefficient about 35.5. But politicization coefficient grows as the number of participants included in the analysis decreases: from 41 for 14 parties to more than 60 for 8 parties.

This indicates, first of all, a sufficiently high level of competition, despite all the subterfuge in favor of United Russia, including ballot stuffing and outright write-ups [28]. The gap between the maximum and effective range is 20-odd percent. This is a lot in absolute numbers, but still far from total fraud. The fluctuation of the politicization coefficient between 40 and 60 points (percent, in fact) suggests that a significant part of those who voted for United Russia did so rather voluntarily (in this case, we are leaving out the integrity of pro-government propaganda methods). Admittedly, the level of this voluntariness is difficult to measure precisely, so we will have to limit ourselves to an approximation of "about half". The fact that the socialization coefficient is always stable and always inferior to the politicization coefficient is most likely an indication of how insignificant the part of the electorate that yields to administrative pressure is. That said, even the 35-36% figure should be halved because it takes into account both those who yield and those who resist.

In any case, the role of direct fraud in securing the victory of the "party of power" should not be absolutized: despite how unfair elections in Russia are, it appears that the government has found a key to the hearts of many Russians, and it is not worth holding out hope that the mass political consciousness will be magically reborn if the regime changes.

As for the second electoral cleavage, it is quite the opposite: its indicators consistently decrease as the number of parties included in the analysis decreases. This applies to both maximum and effective ranges, as well as to politicization and socialization coefficients. This confirms the above-mentioned assumption that this cleavage primarily correlates with voting for "minor" parties and is rather marginal.

On the one hand, the absence of a large gap between the maximum and effective ranges of the second electoral cleavage indicates that election fraud did not take place, i.e. the ballots were cast exclusively for the "party of power" and there was no redistribution of votes among the other participants.

On the other hand, the marked gap between the coefficients of politicization and effective range may be regarded as an indication that only about half of the voters who vote for opposition parties are driven by "ideological" motivations; the rest are guided by some other reasons (for example, "to flip the bird" to the government , which is in fact the essence of "smart voting").

Finally, the low socialization coefficient can be considered as evidence that opposition sympathies barely depend on the social status of voters. This connection cannot, of course, be entirely dismissed. For example, the effects of wellbeing and economic independence are not equal to zero.

However, all these values only indicate the average. They may vary greatly from region to region. The regions of Russia, as noted above, are very different, so the next chapter will consider the structure of the cleavages in the results of the 2021 State Duma election at the regional level.

Electoral Cleavage Structure in the 2021 Election: Regional Level

The 2021 State Duma election was held in all 85 regions, but not all of them had the number of territorial election commissions that submitted cases for calculating factors of territorial differences in party voting and factors of socio-demographic differentiation that would be sufficient for factor analysis. In this regard, the federal subjects with a small number of TECs were merged with bordering regions: Nenets Autonomous Okrug with Arkhangelsk Oblast, Jewish Autonomous Oblast with Amur Oblast, and Sevastopol with the Crimea. The calculations were thus carried out for 82 territorial units.

The results came out extremely diverse, so in order not to drown in detail, the types of political sub-dimensions were grouped by association.

For example, there is a separate category for subject cleavages associated with the new parties.

All varieties of sub-dimensions from the domestic policy issue domain are grouped into one "authoritarian-democratic" sub-dimension (after all, the difference between the government-opposition and loyalists-opposition cleavages is not that considerable); the "authoritarians-democrats" sub-dimension from the foreign policy issue domain are also included here.

Confrontations in the socio-economic domain fall into two categories: "government vs. opposition" and "market liberals vs. social protectionists".

The sub-dimensions of "hawks vs. doves" and "Soviet traditionalists vs. progressives" from the systemic and worldview domains are grouped separately.

As a result, the original 11 sub-dimensions yielded seven: 1) authoritarian–democratic; 2) SE: government–opposition; 3) SE: market liberals — social protectionists; 4) "hawks"–"doves"; 5) Soviet traditionalists–progressives; 6) liberals vs. conservatives in the worldview domain; 7) associated with new parties.

For the sake of simplicity, we will refer to all territorial difference factors as electoral cleavages. Those that have both a political interpretation and a social base will be referred to as complete. Regarding the others, we will point out whether they have only political content or only social content.

Table 14 summarizes the data on the electoral cleavages in the 2021 State Duma election in 82 territorial units. The number of territorial difference factors in party voting varied from two to six. Thirty-eight regions had four ECs each, 22 had three each, 12 had five each, eight had two each, and two had six each. In many regions, the territorial difference factors turned out to be significantly greater in number than at the federal level, a clear indication that intra-regional differences in cleavage structure significantly exceed inter-regional ones.

In the vast majority of regions, the first and second electoral cleavages had either a political content or a social base, or both. In more than half of the cases, third cleavage had either a political interpretation, or a social base, or both.

The most frequent types of political content of ECs were the second political dimension (73 cases), authoritarian–democratic confrontation (70), sub-dimensions related to new parties (67), the first political dimension (64), confrontation between market liberals and social protectionists (59), confrontation between liberals and conservatives in the worldview domain (48), confrontation between government and opposition in the socioeconomic domain (47). In all cases, we were documenting any presence of this or that (sub)dimension in the regression models.

As far as the dominant factors were concerned, more often than not these were the sub-dimensions connected with the new parties (67), followed by authoritarian-democratic confrontation (57), confrontation between government and opposition in the socioeconomic domain (42), market liberals and social protectionists (40), liberals and conservatives in the worldview domain (38).

It is also characteristic that the main political dimensions, authoritarian-democratic and both variations of socioeconomic sub-dimensions, contributed to the first and second electoral cleavages for the most part; the confrontation between liberals and conservatives in the worldview domain amassed in the second and third cleavage; the opposition between Soviet traditionalists and progressives, as well as the sub-dimensions associated with new parties, relatively evenly spread between the first and the fourth ECs.

Among the types of social base of electoral cleavages, urbanization level had the most present (69 cases, dominating in 66), followed by the level of economic activity (41 and 31) and demographic characteristics (39 and 28). The "other factors" category is residual.

Furthermore, urbanization level most often constituted the social base of the first electoral cleavage, while economic activity and demographic characteristics constituted the subsequent ones (especially the second and third).

The relationship between the political (sub)dimensions and the factors of socio-demographic differentiation deserves closer consideration. Table 15 shows all the correlation between them in cases of complete electoral cleavages within the regression models framework (without taking into account the relative weight of each sub-dimension or factor of socio-demographic differentiation).

It is easy to see that urbanization level (337 cases of connections) was the most "active", both in total number of correlations and that in each specific category. Only in the case of the liberals vs. conservatives confrontation in worldview domain did it slightly give way to the factor of economic activity of the population (24 cases against 25). However, urbanization level most frequently correlated with the following categories: "New parties" (56), second political dimension (54), authoritarian-democratic confrontation (42), first political dimension (39), and market liberals and social-protectionist confrontation (34).

The economic activity factor ranked second (254 correlation cases), surpassing its closest competitor — the "Demographic characteristics" factor — in all "nominations" except two: market liberals vs. social protectionists and Soviet traditionalists vs. progressives (they went head-to-head in this case). This factor most often correlated with the following political (sub)dimensions: "New Parties" (44 cases), authoritarian–democratic (38), second political dimension (32), first political dimension and liberals vs. conservatives confrontation in the worldview domain (25 each), "market liberals vs. social protectionists" (24).

The "Demographic characteristics" factor ranked third with 215 cases, overtaking the "Social wellbeing" factor both in total sum and in each individual category. The factors that correlated with it most frequently included "New parties" (38 cases), authoritarian-democratic (31), the second political dimension (28), market liberals vs. social protectionists (24), liberals vs. conservatives on worldview (23), and the first political dimension (21).

Among the political (sub)dimensions, the "socially conditioned" ones often included "New parties" (195 cases), authoritarian-democratic confrontation (153), the second political dimension (151), the first political dimension (113), market liberals vs. social protectionists confrontation (111), liberals vs conservatives confrontation on worldview issues (99).

There should not be any ambiguity about the fact that the "New parties" category took the lead in the overall score as well as in all varieties of socio-demographic differentiation (it was surpassed only once — in the "Social wellbeing" factor — by the authoritarian-democratic confrontation with 18 vs. 19), since it serves as an umbrella for sub-dimensions from all four issue domains.

Speaking of actual leaders, it was the authoritarian-democratic confrontation, especially since it also largely determined the nature of the second political dimension, which had United Russia and Yabloko as the two poles. Variations of the socioeconomic dimension ranked second: the government vs. opposition confrontation and the market liberals vs. social protectionists confrontation. In its pure form, the confrontation between government and opposition in the socioeconomic domain seemed to lag behind, but we should not forget that it was an integral part of the first political dimension. The worldview domain ranked third: Soviet traditionalists vs. progressives (71 cases) and liberals vs. conservatives on worldview issues (99 cases) pushed the confrontation between "hawks" vs. "doves" (42) to the brinks of the systemic issue domain.

Let us now see what combinations of dominant political (sub)dimensions and dominant factors of socio-demographic differentiation occurred most frequently. Table 16 shows the data on these combinations by region (the only missing data is that on the sixth "electoral cleavage", as it occurred in two regions only, and was incomplete in both of them).

It should be noted that complete electoral cleavages occurred in almost all regions (except Astrakhan Oblast and Karachay-Cherkessia). There was one complete EC in 11 regions, two in 33, three in 26, four in eight, and five in two. (That being said, it matters that the cleavages considered here are not institutionalized like those in Lipset and Rokkan's theory, meaning they do not signalize the stability of party systems. Our cleavages include any and all, even the most unstable and situational political confrontations that have at least some social content).

The most frequent combination of political and social content turned out to be the urbanization-based authoritarian-democratic confrontation: 23 out of overall 31 occurred in the first electoral cleavage, three in the second, and four in the third.

With a large gap, it was followed by combinations "new parties / demographic characteristics" and "power vs opposition in socioeconomic issues / urbanization" of 12 cases each. That said, while in SE-GO/Urb 10 out of 12 cases occurred in the first EC, in New/Dem most of the cases occurred in the second and the third (four in each).

Among even rarer combinations were "authoritarian-democratic / economic independence" (10 cases with three occurring in EC-1 and four in EC-2), "market liberals vs. social protectionists / urbanization" (9 cases with four in EC-1 and five in EC-2), "liberals vs. conservatives on worldview issues / other factors" (8 cases with two in EC-1, four in EC-2 and two in EC-3), "second political dimension / urbanization" (8 cases, all in EC-1), etc.

In comparing the 2021 results with those from 2016, interaction between political confrontations (dimensions) and socio-demographic differentiation factors was used as the starting point (Table 17).

First of all, we should point out that the total number of correlations between political and socio-demographic factors was almost 30% less in 2016: 794 vs. 1107 in 2021. This can be explained by the lack of a number of political sub-dimensions at the time: first of all it is the "New Parties" (minus 195 at once), "Soviet traditionalists vs. progressives" (minus 71), and the entire fourth political dimension, largely related to new parties running in elections (minus 50). There were only three main political dimensions in the 2016 election: the first consisted mainly of a confrontation between "hawks" and "doves," the second combined elements of the hawk-dove, authoritarian-democratic and liberal-conservative worldview sub-dimensions, and the third combined liberal-conservative (worldview) and market-protectionist sub-dimensions [19].

Like in 2021, urbanization was the most common out of the socio-demographic differentiation factors: 241 cases, where 56 occurred in the second main political dimension, 53 in the authoritarian-democratic sub-dimension, and 27 in the market liberals vs. social protectionists confrontation.

However, instead of economic independence, the second place was occupied by demographic characteristics: 218 cases, where 56 occurred in the second main political dimension and 48 in the authoritarian-democratic confrontation.

The economic independence factor came in third place, and with a large gap at that: 162 cases, where 35 correlated with the authoritarian-democratic sub-dimension and 24 with the second main political dimension.

The authoritarian-democratic confrontation beat all others in terms of socio-demographic differentiation factors with 171 cases. The second political dimension came as a close second (165 cases), followed by the third PD with a large gap at 96 cases. Next came the confrontations: market liberals vs. social protectionists (88 cases) and government vs. opposition on socioeconomic issues (86 cases).

As for the dominant combinations of political and social factors, table 18 contains data on each region and each electoral cleavage, except for the sixth one which yet again occurred in two regions only (Kabardino-Balkaria and Kostroma Oblast). That said, it was a complete one in the second case ("government vs. opposition on socioeconomic issues / other factors").

In 2016, only one region — Karachay-Cherkessia — indicated no complete ECs whatsoever; 10 regions indicated one complete EC each, 40 indicated two each, 28 indicated three each, two (Kostroma and Novgorod Oblasts) indicated four each, and one (Karelia) indicated five.

Like in 2021, in 2016 the most frequent combination of political and social factors was that of authoritarian-democratic confrontation and the level of urbanization. There were eight more such cases: 39 compared to 31 in 2021, and 36 of them (compared to 23 in 2021) occurred in the first electoral cleavage.

On the other hand, the second most common combination in 2016 was that of "market liberals vs. social protectionists / urbanization": 22 cases (in 2021 it ranked fifth with 9 cases). That said, in 13 regions it occurred in the second electoral cleavage and in 7 regions in the first (in 2021 the same ratio was 5 to 4).

The authoritarian-democratic confrontation with demographic characteristics as a social base ranked third in 2016: 14 cases (seven in the first electoral division, three in the second). In 2021, there were only four such cases.

Then came combinations of "second political dimension / urbanization" (10 cases, all in the first electoral cleavage), "government vs. opposition on socio-economic issues / economic independence" (9 cases spread between the second and fifth electoral cleavage), "authoritarian-democratic / economic independence" (8 cases, including 7 in the first EC), etc.

In other words, 2016 was less diverse: certain political sub-dimensions were absent altogether (primarily those related to new parties that had not yet emerged), but the concentration of dominant types in the first and second electoral cleavages was somewhat higher.

The share of the "authoritarian-democratic opposition / urbanization" combination had decreased by 2021, although not as dramatically as that of the "market liberals vs. social protectionists / urbanization" combination. On the other hand, in 2021 the "government vs. opposition on socio-economic issues / urbanization" combination was more common, rising from 8 to 12 cases. Nearly all of them amassed in the first electoral cleavage (in 2016 they spread between the first and the third).

And, of course, the 2021 election was characterized by a host of new sub-dimensions associated with the new parties. A significant part of these sub-dimensions correlated with the demographic characteristics factor (but not that one alone — together with the other factors, their number amounted to 32). The confrontation between Soviet traditionalists and progressives, which did not exist at all in 2016, also emerged for the first time (for more on the role of the Soviet-past agenda in the 2021 campaign, see [21]).

Now let us see what happened to the trend of blurred structure of electoral cleavages that was discovered when comparing the results of the 2011 and 2016 State Duma elections [17]. By comparing the number of factors of territorial differences in party voting and full electoral cleavages in 2021 and 2016, we can conclude that this was not a confirmed trend to say the least, and perhaps it was even reversed.

The total number of territorial difference factors in the 2021 party voting decreased by 11 compared to 2016, but the number of complete electoral cleavages, on the other hand, increased by 10.

In 21 regions, the number of territorial difference factors decreased (by one in 17 regions, by two in three, by four in Kabardino-Balkaria — from six to two). In 13 regions, the number increased (by one in 10, by two in three). In 48 regions, the number remained unchanged.

The number of complete electoral cleavages increased in 30 regions (by one in 25, and by two in five), decreased in 19 (by one in 13, and by two in six), and remained unchanged in 33. The relative increase in the number of complete electoral cleavages was largely due to new sub-dimensions emerging with the new parties.

Another important aspect is that in the vast majority of complete electoral cleavages, the politicization and socialization coefficients are usually relatively low, and therefore cover a small part of the electorate. It is not uncommon for a high politicization coefficient to come with a low socialization coefficient of or vice versa. This means that the electorate's increased awareness of the essence of the political conflict is not always socially conditioned, and on the contrary, the obvious social content of the vote is not always related to the political preferences of voters.

We shall therefore consider how many regions in the 2016 and 2021 elections display electoral cleavage with high (≥30) politicization and socialization coefficients, which indicate a high level of political competition resting on a solid social base. (Incidentally, the author has never come across cases when several electoral cleavages with high politicization and socialization coefficients occurred in the same region during the whole period. There was either one dominant cleavage, accompanied by one or more smaller ones, or several cleavages that were roughly equal in scope.)

Table 19 shows the corresponding regions, the ratio of politicization and socialization coefficients of major electoral cleavages (Pc/Sc), as well as the dominant content of the related political (sub)dimensions and factors of socio-demographic differentiation (Pol/Soc). In almost all cases, these ranked first in the hierarchy of electoral cleavages; only in Astrakhan Oblast (2016), Khakassia, and Sakhalin Oblast (both 2021) did they rank second.

Let us do some calculations. While in 2016 electoral cleavages with high politicization and socialization coefficients were recorded in 21 regions, in 2021 the number increased to 30. Thirteen regions indicated these ECs in both 2016 and 2021. These are, one might say, regions with a tradition of intense political competition, which a solid social base. Most of them are located in the European part of Russia: Belgorod, Voronezh, Ivanovo, Kaliningrad, Lipetsk, Oryol, Penza, Ryazan, Samara, Tver, and Ulyanovsk Oblasts. There are some exceptions, too, like the Republic of Adygea and Magadan Oblast. In the 1990s, mosty of these regions (with the exception of the Ivanovo, Kaliningrad, Samara, Tver, and Magadan Oblasts) belonged to the Red Belt — a group of regions which gave stable support to the Communist Party.

The 2016 and 2021 results do not differ in quantity alone. The most notable difference is the diversity of political content of electoral cleavages. In 2016, 19 out of the 21 cases represented authoritarian-democratic confrontations, and considering that the same confrontation formed the basis of the second political dimension, it may have well been 20 cases. And only in one region — in Astrakhan Oblast — it was a confrontation between market liberals and social protectionists (the only special case in 2016, when the electoral cleavage with high politicization and socialization coefficients ranked second in the hierarchy).

In 2021, only 10 (including the second political dimension - 15) out of 30 electoral cleavages represented authoritarian-democratic confrontations. Seven had to do with the socioeconomic issue domain: in two cases it was a confrontation between market liberals and social protectionists, and in five cases it was a confrontation between government and opposition. Four electoral cleavages correlated with the first main political dimension, which combined elements of the "hawks" vs. "doves" confrontation, as well as the government vs. opposition confrontation on domestic policy and socioeconomic issues. One was correlated only with the "hawks" vs. "doves" confrontation and three correlated with new parties.

Regarding the social base, in 2016, it was urbanization for 15 cases out of 21, demographic characteristics in three, economic independence in two, and social wellbeing in one. In 2021, urbanization presented in 25 of 30 cases; the remaining five fell into the demographic characteristics category. The latter circumstance runs counter to the previously noted tendency to displace the demographic characteristics factor by that of economic independence as a social basis for political confrontations. It is clearly too early to discard demographic characteristics, that is, if we are talking about highly competitive regions and not the average.

Conclusion

Study results indicate a fairly high level of "fluidity" in the structure of electoral cleavages. It is hardly a trend that occurs only in Russia. If one looks at them up close instead of from afar, it is likely the cleavages will turn out that way everywhere.

The political content of ECs is particularly fluid, while the social base is more inert. This is hardly a surprise since the political side is handled by active players: parties whose task is to come up with a proposal and put the "product" on the market, while voters act as buyers, approving or disapproving of the new "product".

Any hope that the concept of cleavages could help explain the sustainability of party systems may have been vain to begin with. Even if we disregard the fluidity of the political content of cleavages (which Lipset and Rokkan clearly had good reason to ignore), the relative inertia of social interest conflicts cannot be recognized as a sufficient basis for such stability. First, the social structure of any society does change, even if slowly. Second, a voter belongs not to one, but several social groups, meaning the situation it is more advantageous to appeal to him or her as a representative of a certain class or stratum, ethnicity, religious denomination, etc.

That said, the party "market" is ruled by fierce competition, which is impossible to get rid of even if one monopolizes a certain part of it — by creating preferences for certain players, for example — as the competition will simply move to the remaining part of the political space.

It would seem, therefore, that cleavage studies should not focus solely on constant features of cleavages, but rather on changes in their structure. Studying these changes is the key to identifying the constant, not the other way around.

This study recorded the changes that occurred over a relatively short time period between 2016 and 2021 (or between 2011-2021 in we consider the author's previous works). This is quite enough to state that factors of interregional sociodemographic differentiation affect voter attitudes not individually, but in a complex manner: each of the identified factors affects nearly every political confrontation, just to a different extent. Urbanization level affects them more, social well-being of the population does so less, and such factors as demographic characteristics and economic independence compete for second place. If one wants to know whether these fluctuations indicate some long-term trend or are just temporary, longer observations will be required.

So far, all that is clear is that the factors act as a cluster. Individually, they are something like shadows on the wall of Plato's cave. Whatever "entity" they reflect may become clear with the passing years.

However, to claim that these factors produce certain political confrontations would be an obvious sociologism. Rather, the opposite is true: active political actors — parties and government elites — fight among themselves and generate political confrontations expecting that the latter will resonate with some social contradictions and bring dividends represented by extra votes in elections. In the context of Russian electoral authoritarianism (where "authoritarian" has been consistently pushing out the "electoral" in recent years), the leading role in creating such confrontations belongs, however, not to the parties, but to the governing elite.

Over the past two decades, it was the government that generated the confrontations that resonated with the masses. First of all, it is the authoritarian-democratic confrontation, which emerged following centralization of administrative resources that took place between late 1990s and early 2000s. It found its social base in the differentiation between those voters who were able to resist the administrative pressure and those who were not.

Another confrontation generated by the governing elite is the cleavage between "imperialists" and "anti-imperialists" ("hawks" and "doves"), which came to the fore in 2012-2014. Before that, Communists and Zhirinovsky supporters had spent decades pushing for a tougher foreign policy, but their efforts had not been particularly successful. When Vladimir Putin had returned to the Kremlin, he became alarmed by the "white ribbon" protests (2011-2012) and moved to take over the leftist-conservative agenda, first by pushing through laws on foreign agents and undesirable organizations, and then by interfering in Ukrainian affairs. It is precisely then when this confrontation ceased to be marginal and became a core one.

The same can be said about the confrontation between Soviet traditionalists and progressives, which had a serious impact on the political space in the 1990s and early 2000s, yet faded out almost completely by the 2007 election [22]. In the course of preparations for the 2021 State Duma election, it was the pro-government actors, and not the Communists, who introduced the Soviet-past agenda to inter-party debate (although the Communists never really abandoned it, simply lacking the resources for promotion).

Newly emerged parties may have spiced up the inter-party debate, but they were unlikely to have done it without assistance from the Presidential Executive Office. After all, it falls on them to decide who can and who cannot run in elections. The new parties were clearly tasked with picking up the part of the electorate that were out of reach of the old ones.

At any rate, an analysis of the changes in the cleavage structure in the 2021 State Duma elections at the federal and regional levels suggests that the political elites (the government above all) have done a good job "probing" the electorate for any weak spots.

Without denying the role of direct electoral fraud in securing the dominant position of the "party of power," it should be recognized, however, that the main bet was not on them, but on the specific work with the mass political consciousness (that is, monopolizing the media landscape and consistently curtailing freedom of speech).

One can hardly call these endeavors unsuccessful. They proved quite effective on a large part of the electorate — on its less urbanized, less economically independent, older, and perhaps poorer part.

Fluid political and social content of electoral cleavages do not, however, prevent certain constants from reoccurring. The authoritarian-democratic confrontation remains the leading political one with the most solid social base. Like before, from a social standpoint it primarily correlates with urbanization level, but at the same there is also correlation with the population's demographic structure and economic independence.

The second place still belongs to confrontations on socioeconomic issues: between government and opposition as well as between market liberals and social protectionists. The former is gradually gaining momentum, while the latter is losing, perhaps due to being pushed out by confrontations on foreign policy and worldview isues ("hawks" vs. "doves," conservatives vs. liberals, Soviet traditionalists vs. progressives). Urbanization level is also the primary social base for socioeconomic confrontations, but so are (in part) demographic characteristics and economic independence. That said, it is difficult to definitively assess the outcome of the competition between the two: economic independence prevails based on average regional values, and demographic characteristics prevail based on values from the most competitive regions.

The novelty of the 2021 elections is the expanded range of political confrontations that inform electoral cleavages. The authoritarian-democratic confrontation seems to have lost its overwhelming dominance. It was diluted by the usual competitors from the socioeconomic subject area, as well as rather exotic ones, including those associated with the emergence of new parties. Perhaps this explains the trend toward blurred electoral cleavage structures, which emerged in the comparative analysis of the 2011 and 2016 election results [17]. In 2021, the total number of the ECs decreased, but the number of complete ECs, including those characterized by high politicization and socialization coefficients, increased.

It is difficult to say whether this trend is persistent or not, especially after February 24, 2022, when the situation in the country changed dramatically.

In any event, observing and documenting changes in political content and social base of electoral cleavages requires a lot of time and effort, especially if one wants to move on to higher-level generalizations.

Received 31.03.2023, revision received 27.04.2023.

---

References

1. Akhremenko A.S. Struktura elektoralnogo prostranstva [Electoral Space Structure]. – Moscow: Sotsialno-politicheskaya mysl, 2007. 320 p. (In Russ.)
2. Albright J.J. The Multidimensional Nature of Party Competition. – Party Politics. 2010. Vol. 16. No. 6. P. 699–719.
3. Bornschier S. Cleavage Politics and the Populist Right: The New Cultural Conflict in Western Europe. Philadelphia: Temple University Press, 2010.
4. Budge I. The internal analysis of election programmes. – Ideology, Strategy and Party Change: Spatial Analyses of Post-War Election Programmes in 19 Democracies, Budge I., Robertson D., Hearl D. (eds.). Cambridge: Cambridge University Press, 1987. P. 15–38.
5. Franklin M. The decline of cleavage politics. – Electoral change: Responses to evolving social and attitudinal structures in Western countries, edited by M. Franklin, T. Mackie, and H. Valen. Cambridge, UK: Cambridge University Press, 1992. P. 381–402.
6. Häusermann S., Kriesi H. What do voters want? Dimensions and configurations in individual-level preferences and party choice. – The Politics of Advanced Capitalism. Cambridge: Cambridge University Press, 2015. P. 202–230.
7. Hooghe L., Marks G. Cleavage theory meets Europe’s crises: Lipset, Rokkan, and the transnational cleavage. – Journal of European Public Policy. 2018. Vol. 25. No. 1. P. 109–135.
8. Hooghe L., Marks G., Wilson C.J. Does left/right structure party positions on European integration? – Comparative Political Studies. 2002. Vol. 35. No. 8. P. 965–989.
9. Huber J., Inglehart R. Expert Interpretations of Party Space And Party Locations In 42 Societies. – Party Politics. 1995. Vol. 1. No. 1. P. 73–111.
10. Inglehart R. Cultural shift in advanced industrial society. Princeton, NJ: Princeton University Press, 1990.
11. Janda K. A Tale of Two Parties: Living Amongst Democrats and Republicans Since 1952. New York and London: Routledge, Tailor and Francis Group, 2021.
12. Janda K. The Republican Evolution: From Governing Party to Anti-Government Party (1860-2020). New York: Columbia University Press, 2022.
13. Kitschelt H., Mansfeldova Z., Markowski R., Toka G. Post-Communist Party Systems: Competition, Representation, and Interparty Cooperation. Cambridge: Cambridge University Press, 1999.
14. Korgunyuk Yu. Cleavage Theory and Elections in Post-Soviet Russia. – Perspectives on European Politics and Society. 2014. Vol. 15. No. 4. P. 401–415.
15. Korgunyuk Yu.G. Kontseptsiya razmezhevanii i faktornyi analiz [Cleavage Theory and Factor Analysis]. – Politeia. 2013. No. 3. P. 31–51. (In Russ.)
16. Korgunyuk Yu.G. Kontseptsiya razmezhevaniy i teoriya problemnykh izmereniy: tochki peresecheniya [Intersections of Cleavages Concept and Issue Dimensions Theory]. – Polis. 2019. No. 6. P. 95–112. DOI: 10.17976/jpps/2019.06.08 (In Russ.) - https://www.politstudies.ru/files/File/2019/6/Polis-6-2019-Korgunyuk.pdf
17. Korgunyuk Yu.G. New Instruments for Measuring Electoral Cleavages and Regional Map of Cleavages During the 2011 and 2016 Elections. – Electoral Politics. 2019. No. 2. - http://electoralpolitics.org/en/articles/novye-instrumenty-izmereniia-elektoralnykh-razmezhevanii-i-regionalnaia-karta-razmezhevanii-na-vyborakh-2011-i-2016-gg/
18. Korgunyuk Yu.G. New Instruments for Measuring Electoral Cleavages: From Macro- to Micro-Level. – Electoral Politics. 2019. No. 1. - http://electoralpolitics.org/en/articles/novye-instrumenty-izmereniia-elektoralnykh-razmezhevanii-ot-makro-k-mikrourovniu/
19. Korgunyuk Yu.G. Politicheskiye izmereniya i predmetnyye razmezhevaniya: metodika opredeleniya vzaimodeistviya [Political Dimensions and Subject Cleavages: A Methodology for Determining Interaction]. – Politeia. 2021. No. 3. P. 147–171. DOI: 10.30570/2078-5089-2021-102-3-147-171 (In Russ.) - http://politeia.ru/files/articles/eng/Politeia-2021-3(102)-147-171.pdf
20. Korgunyuk Yu.G. Structure of Electoral Cleavages in Russia's Regional Parliament Elections (2016–2020): Transformation Trends. – Electoral Politics. 2021. No. 1 (5). - http://electoralpolitics.org/en/articles/struktura-elektoralnykh-razmezhevanii-na-vyborakh-v-regionalnye-sobraniia-subektov-rf-2016-2020-tendentsii-transformatsii/
21. Korgunyuk Yu.G. Tematika sovetskogo proshlogo v dumskoi kampanii 2021 goda [Soviet Past as a Subject Matter in the 2021 State Duma Election Campaign]. – Politeia. 2022. No. 3. P. 105–129. DOI: 10.30570/2078-5089-2022-106-3-105-129 (In Russ.) - http://politeia.ru/files/articles/eng/Politeia-2022-3(106)-Pages-001-202-105-129.pdf
22. Korgunyuk Yu.G. Tematika sovetskogo proshlogo v izbiratelnykh kampaniyakh postsovetskoi Rossii [Soviet Past as a Subject Matter in Election Campaigns in Post-Soviet Russia]. – Politeia. 2023. No. 1. P. 88–125. DOI: 10.30570/2078-5089-2023-108-1-88-125 (In Russ.) - http://politeia.ru/files/articles/eng/Politeia-2023-1(108)-Pages-001-210-88-125.pdf
23. Korgunyuk Yu.G. Vybory po proportsionalnoi sisteme kak massovyi opros obshchestvennogo mneniya [Proportional Elections as Mass Opinion Poll]. – Politicheskaya nauka. 2017. No. 1. P. 90–119. (In Russ.)
24. Kriesi H., Grande E., Dolezal M., Helbling M., Hoglinger D., Hutter S., B. Wuest. Political Conflict in Western Europe. Cambridge: Cambridge University Press, 2012.
25. Lachat R. Which Way From Left To Right? on The Relation Between Voters’ Issue Preferences and Left–Right Orientation in West European Democracies. – International Political Science Review. 2018. Vol. 39. No. 4. P. 419–435.
26. Lawson K. Razmezhevaniya, partii i izbirateli [Cleavages, Parties and Voters]. – Political Science (RU). 2004. No. 4. P. 50–55. (In Russ.)
27. Lipset S.M., Rokkan S. Cleavage Structures, Party Systems, and Voter Alignments: An Introduction. – Party Systems and Voter Alignments: Cross-National Perspectives. New York; London: The Free Press. Collier-MacMillan limited, 1967. P. 1–64.
28. Lukyanova E.A., Poroshin E.N., Arutyunov A.A., Shpilkin S.A., Zvorykina E.V. Vybory strogogo rezhima. Kak rossiiskiye vybory stali nevyborami i cho s etim delat? Politiko-pravovoye issledovaniye s elementami matematiki. [Maximum Security Elections: How Did Elections in Russia Turn Into Non-Elections and What is to Be Done About It? A Political and Legal Research With Elements of Mathematics]. Moscow: Mysl, 2022. 400 p. (In Russ.)
29. Rovny J., Polk J. New Wine In Old Bottles: Explaining the Dimensional Structure of European Party Systems. – Party Politics. 2019. Vol. 25. No. 1. P. 12–24.
30. Rovny J., Whitefield S. Issue Dimensionality and Party Competition in Turbulent Times. – Party Politics. 2019. Vol. 25. No. 1. P. 4–11.
31. Sata R., Karolewski I.P. Caesarean politics in Hungary and Poland. – East European Politics. 2020. Vol. 36. No. 2. P. 206–225.
32. Stoll H. Social Cleavages, Political Institutions and Party Systems: Putting Preferences Back into the Fundamental Equation of Politics. A Dissertation Submitted to the Department of Political Science and the Committee on Graduate Studies in Partial Fulfilment of the Requirements for the Degree of Doctor of Philosophy. – Santa Barbara: University of California, 2004. URL: http://www.polsci.ucsb.edu/faculty/hstoll/research/thesis.pdf (accessed 08.03.2023). - http://www.polsci.ucsb.edu/faculty/hstoll/research/thesis.pdf
33. Teney C., Lacewell O.P., de Wilde P. Winners and losers of globalization in Europe: attitudes and ideologies. – European Political Science Review. 2014. Vol. 6. No. 4. P. 575–595.
34. Torcal M., Mainwaring S. The Political Recrafting of Social Bases of Party Competition: Chile, 1973–95. – British Journal of Political Science. 2002. No. 33. P. 55–84.