# Structure of Electoral Cleavages in Russia's Regional Parliament Elections (2016–2020): Transformation Trends

#### Abstract

Comparing electoral cleavages in the Russian regions at the 2016 State Duma election and elections to regional legislative assemblies in 2016—2020 results in the conclusion that there is a prevalent tendency where the structure of said cleavages is blurred. Even in 2016, this trend manifested itself in over half of the regions, although almost a third of cases indicated consolidation of the structure of electoral cleavages. In the three years that followed, the trend of blurred electoral cleavages only became stronger: only two out of 33 regions indicated consolidation, while seven more managed to maintain a certain status quo. To some extent, this is explained by denying election registration to liberal parties. In 2020, however, more than a third of the regions indicated electoral cleavages consolidating their structure, despite the fact that neither Yabloko nor PARNAS were allowed to participate in the elections there. At the same time, the participants were joined by parties created just before the election – all of which demonstrated clear political leanings. This confirms the hypothesis that administrative regulation of election entry list either clouds or, on the contrary, clarifies the political picture in the minds of voters. Eliminating one of the opposition poles tends to blur the picture. When new challengers to the same niche enter the picture, it becomes clearer. In addition, in different regions of the country, mass political consciousness is at different development levels. The confrontation between the "pro-market" government and the anti-market left and nationalist opposition takes the most complete forms in the former Red Belt regions while becoming more and more blurred in others, as people cease to perceive the previous political landscape and do not see any alternative. In Saint Petersburg, with its rather politically savvy electorate, political landscape is entering a new stage of development.

So far, few studies have been dedicated to mass political consciousness, which is no wonder, as its structure is quite complex. The consciousness of a politically active minority (we will refer to it as “the elite” for convenience sake) is difficult and multidimensional, but it has a certain integrity, albeit established by contradictions. Mass political consciousness, on the other hand, is fragmented. The active minority shapes it into something more or less structured by striking keynotes.

Political actors – government officials, parties, candidates – are fighting among themselves while focusing on the electorate and trying to catch popular sentiments and turn them to their advantage. Voters, in their turn, express their moods by giving preference to actors whose language is more familiar and understandable to them.

The concept of cleavages [17] stems from the assumption that the sentiments of voters is largely determined by their – voters' – social status. According to the creators of the concept, dominant cleavages, which are institutionalized in party confrontations, are based on conflicts of public interest. Seymour Martin Lipset and Stein Rokkan proposed a pattern of possible changes in cleavages in Europe in the 19th–first half of the 20th century, naturally sketching it with broad strokes and simplifying the real state of affairs in many ways.

In essence, they took only the most obvious cleavages into account. The latest use of their tools showed that the scheme they built was applicable to a limited geographic area (northern part of Western Europe) and a short historical period (early 19th – mid-20th centuries). The most important thing is that it turned out that what looks like a single ridge from the skies turns into a complex system of hills and crevices on the ground. The cleavages marked by Lipset and Rokkan turned out to be an interweaving of overlapping cleavages. The former thus simply coexist with the latter and do not fit into the general landscape.

The picture of the dominant cleavages being replaced turned out to be no less complicated. Giving way to the "new" cleavages as the new dominant ones, the "old" cleavages were in no hurry to leave the political scene. For example, in Great Britain, the Labour Party transforming into the main rival of the Conservative Party did not eliminate the latter's opposition with liberals, who, in their turn, still have a sizeable representation in parliament. This means that the conservative-liberal opposition still exists.

To an even greater extent, this blur is inherent in developing democracies, and even more so in electoral autocracies. Russian academic literature [1; 9; 8] has already mentioned that in post-Soviet Russia, the change in the dominating cleavages that took place on the cusp between the 20th and 21st centuries did not make the alignment of forces simpler. While the authoritarian-democratic cleavage pushed the socio-economic cleavage aside, it did not abolish it, and the latter continues to manifest, albeit in the background.

The situation became even more confusing after the 2014 Crimea crisis. The cleavage hierarchy in electoral space remained unchanged (authoritarian-democratic first, socio-economic second), but in political space, that is, the sphere of "elite" consciousness, the opposition between "imperialists" and "anti-imperialists" came to the foreground. In a way, this opposition may be considered a systemic cleavage that manifested as the rivalry between those who favor openness/closedness of social systems [14; 12]. Although the influence this change had on the mass consciousness was only indirect, it nevertheless obscured the general picture significantly.

In some cases, this openness/closeness is temporal, while in others it is spatial. In the former case the opponents focus on the future/past, while in the latter they focus on supranational organizations integration or, on the contrary, isolationism. The "temporal" variation can be observed in the opposition between materialists and postmaterialists [7], new and old politics [3; 18]. The "spacial" variation can, in its turn, be observed in the opposition between demarcationists and integrationists (i.e. those who benefited from globalization and those who did not) [16], universalists and particularists [4], cosmopolites and communitarianists [21]. There are also combined variations that oppose libertarian-universalists to traditionalist-communitarianists [2] or  Green/alternativists/libertarians to traditionalists/authoritarians/nationalists (GAL/TAN) [6; 5].

The author's previous publication [11] pointed out that from a voter's perspective, the political picture of the 2016 State Duma election tends to be more complicated as compared to 2011, even in the most uncompetitive regions of the Russian Federation. As for the most competitive regions, the political picture tends to be blurred. However, what would be interesting is to compare the results of the Duma elections with those of regional legislative assembly elections from this point of view.

Of course, such a comparison may not seem entirely correct, but it at least allows for preliminary conclusions to be drawn. The task is made somewhat easier by the fact that in 2016, in 39 regions of Russia, elections to regional legislative assemblies were held simultaneously with the Duma election, which makes the issue of correctness irrelevant in their regard. As for other regions, the situation is more complicated, but general trends can be traced in one way or another.

Under a proportional representation system, the main difference between elections to the federal parliament and those to regional legislative assemblies is that in the Duma campaign, voters throughout the country are provided with the same set of participating parties (another thing is that one way or another, the electorate reduces this set for itself in different regions, for more details, see [Korgunyuk 2019b]). As for elections to regional legislative assemblies, the number of party participants is set by the regions themselves, and is never higher than in the Duma elections – however, it is typically set to at least for four. In theory, this should simplify the structure of electoral cleavages and make it lend itself to political interpretation more easily as well as put it deeper in social context.

On the other hand, parties that are unable to participate in federal elections often do so in regional elections, including the freshest newcomers, whose positions on a significant part of issues are unknown even to experts (and the newcomers themselves), not to mention voters. This makes the political picture more confusing for the latter and, in theory, should also erode the structure of electoral cleavages.

The study is going to reveal which of the tendencies prevails.

## Research methodology

To a certain extent, the proposed methodology repeats the one the author used in his previous papers [12; 11]. Let us run through its basics.

The starting point is the factors of territorial dispersion of votes cast for different parties under proportional representation system. These factors are calculated by means of factor analysis of the percentages of votes received by parties in different territorial units; parties are variables, while territorial units are cases. If these factors have a political interpretation and social connotations, then they are recognized as electoral cleavages.

Whether factors can be interpreted politically is determined by comparing factor loadings of parties in electoral and political spaces using correlation and regression analyses. Factor loadings in the political space are calculated based on the parties' positions on the most polarizing and controversial issues (documents and materials describing party positions can be found in the PartArchive database, http://www.partinform.ru/pa98). Positions of parties are evaluated on a scale from –5 to +5 and subjected to factor analysis, where the parties are variables and the issues are cases.

The author previously defined the resulting factors as political cleavages [9; 14; 8] while setting them apart as follows: 1) socio-economic cleavage, which was considered dominant in Western countries until the end of the 20th century; 2) authoritarian-democratic cleavage, which was not very relevant in the West, but quite widespread in post-Soviet countries and Latin America, and which has been resurfacing in Central and Eastern Europe recently [1; 22; 20: 32; 19]; 3) systemic cleavage, which was mentioned above.

However, the author's point of view changed over time, and he now considers these factors more like dimensions where political cleavages are stored (for more details, see [10; 15]). Sometimes the dimensions themselves can act as cleavages, but, as a rule, mass consciousness is unable to grasp the political space completely, only snatching separate pieces from it.

In order to take this feature into account, the author decided to distribute all the questions into three main issue domains – domestic political, socio-economic and systemic (international + ideological) – and subject each area to factor analysis individually. As a result, each of them revealed two or three "object cleavages".

For example, in the domestic political issue domain, these aspects are mostly as follows: 1) the confrontation between the government and the rest of the electoral participants (we shall refer to it as the government-opposition variation of authoritarian-democratic cleavage – AD-GO); 2) the confrontation liberals and loyalists (AD-LiLo) or liberals and communists (AD-LiC), or loyalists and the opposition (AD-LoOp).

In the socio-economic issue domain, these aspects are: 1) the government vs. the rest (SE-GO); 2) liberals vs. the communists (SE-LiC) or social paternalists vs. market-oriented individuals (SE-SPM). These subtle aspects manifest when liberal parties – primarily Yabloko – are either registered for an election or denied registration.

In the systemic issue domain, these aspects are: 1) imperialists vs. anti-imperialists (Syst-IAI); 2) Soviet traditionalists vs. progressives (Syst-STP) or Soviet traditionalists vs. loyalists (Syst-STLo). In this case, it is the same.

Extra cleavages may appear in each issue domain. In regional elections, this most often depends on the number and composition of participants.

To find out whether the factor of territorial dispersion has a political interpretation, the factor loadings of parties in the electoral space are compared with their factor loadings within political dimensions and issue domain cleavages. For this purpose, it is better that correlation analysis is replaced with multiple regression, where factor loadings in the electoral space are a dependent variable, and factor loadings in political dimensions and issue domain cleavages are independent variables. If the statistical significance of the coefficients in the resulting regression model does not exceed 0.1, we can assume that we are dealing with an electoral cleavage that has a political interpretation.

To find out whether these electoral cleavages have a social connotation, multiple regression is used as well, but this time the dependent variable is the factor scores of territorial units of the factors of inter-territorial dispersion (electoral cleavages). In turn, demographic and socio-economic indicators of the same territorial units are used as independent variables and are preliminarily subjected to factor analysis (the basic factors of inter-territorial differentiation that were identified include the quality of life level that nearly coincides with the urbanization level; demographic characteristics; the level of economic independence of the population and territories; the level of social wellbeing, etc.; see [14; 12; 11]). If the link between the regression model and the statistical significance of the coefficients is less than 0.1, electoral cleavages are considered to have a social connotation.

At the same time, the author does not use the beta coefficients of regression models to directly describe the structure of electoral cleavages (EC) in elections to regional legislative assemblies, but opts for the measurement tools proposed in his previous works instead [12; 11]:

maximum range coefficient formula: $$Mc = \Sigma i |FLi|$$, where $$i$$ is the share of votes received by each party under the proportional representation system and $$|FLi|$$ is the factor loading 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 formula: $$Pc = Ec R^2$$, where $$R$$ is the multiple regression coefficient (coefficient of determination) of the relationship between the EC and a set of political dimensions and issue domain cleavages;

– 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 that 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 results are conditioned by the social status of voters.

Thus, 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, while a low value indicates that they take little interest in it. In the same way, 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.

However, we shall focus on how "complete" the electoral cleavages turned out to be in each of the cases instead of on the values of said coefficients. In this regard, all regions are divided into the following types (based on the structure of the EC):

1) "complete" – all electoral cleavages have both a political interpretation and a social connotation;

2) "rich" – at least two electoral cleavages have both a political interpretation and a social connotation;

3) with one "complete" electoral cleavage that ranks first in the hierarchy; variations are possible in this case: subsequent factors of territorial dispersion have either a political interpretation or a social connotation, or they do not have either;

4) with at least one "complete" electoral cleavage that does not rank first in the hierarchy; variations are possible here as well, like in the previous case;

5) with "incomplete" electoral cleavages, that is, having either only a political interpretation, or only a social connotation; variations are also possible.

At the same time, attention is paid to the trend in changes of the structure of electoral cleavages as compared to the State Duma elections, as in whether there was consolidation or, on the contrary, erosion. Consolidation implies that the number of ECs decreases, they become "complete" and coefficients of effective range, politicization and socialization increase. Erosion implies trends that are completely opposite.

## 2016–2020 regional legislative assembly elections in the context of the changing structure of electoral cleavages

Table 1 summarizes the results of regression analysis (multiple linear regression; ordinary least squares method) of links between electoral cleavages with political dimensions and issue domain cleavages. The list does not include Moscow, where elections to the City Duma (2019) were held only in single-seat constituencies (and United Russia candidates ran as independents at that), and the Nenets Autonomous Okrug, which, as was mentioned above, was combined with the Arkhangelsk Oblast in the calculations. In a total of 83 regions, the author identified 203 regional dispersion factors in voting for parties. Notably, 9 regions indicated four such factors each, 26 regions indicated three, 41 regions indicated two , and 7 regions indicated only one. Only 97 of these 203 factors had a political interpretation; 15 of them with a statistical significance of less than 0.1, and the remaining 82 with a significance less than 0.05.

###### Table 1. Regression models of connection between electoral cleavages and political dimensions (issue domain cleavages)
 Region EC R2 Political dimension - 1 Political dimension - 2 Political dimension - 3 Authoritarian-democratic -1 (Government - Opposition) Authoritarian-democratic - 2 Liberals - Loyalists, Liberals - Communists, Loyalists - Communists) Socio-Economic -1 (Government - Opposition) Socio-Economic -2 (Social paternalists - Market supporters) Systemic -1 (Imperialists -Anti-imperialists) Systemic - 2 (Soviet traditionalists - progressives) 2016 Republic of Adygea EC-1 0.976 0.988 (0.090) Republic of Mordovia EC-1 0.991 –0.996 (0.067) Chuvash Republic EC-1 0.998 –0.327 (0.049) –1.384 (0.040) –0.521 (0.044) Altai Krai EC-1 0.697* 0.835 (0.318) Krasnoyarsk Krai EC-1 0.559* –0.521 (0.258) –0.668 (0.258) EC-2 0.505 –0.711 (0.249) Perm Krai EC-1 0.611* 0.782 (0.312) Primorsky Krai EC-1 0.691 0.831 (0.278) Stavropol Krai EC-1 0.616* 0.785 (0.310) EC-2 0.883 –0.508 (0.200) –0.874 (0.200) Amur Oblast EC-3 0.504 –0.710 (0.266) Astrakhan Oblast EC-1 0.612* 0.782 (0.312) Vologda Oblast EC-1 0.610 –0.781 (0.279) Kaliningrad Oblast EC-1 0.750 0.866 (0.204) Kamchatka Krai EC-1 0.988 –0.881 (0.079) 0.347 (0.079) EC-2 0.978 –0.481 (0.107) –0.938 (0.107) Kursk Oblast EC-1 1 1.939 (0.022) 0.935 (0.029) –0.243 (0.011) Lipetsk Oblast EC-1 0.977 1.180 (0.091) –0.664 (0.091)

The table is not fully displayed Show table

Only electoral cleavages with the determination coefficients significant at p<0.05 (*p<0.1) are shown.
Electoral cleavages without political interpretation were detected in the following regions:
2016 - Republic of Dagestan (2 EC), Republic of Ingushetia (2), Republic of Karelia (3), Chechen Republic (2), Kirov Oblast (2);
2018 - Republic of Khakassia (3), Republic of Sakha (Yakutia) (1);
2019 - Altai Republic (4), Volgograd Oblast (1);
2020 - Kurgan Oblast (2).

The most common variation of political content of electoral cleavages (manifesting either alone or in combination with other variation) turned out to be the government-opposition variation of socio-economic cleavage (SE-GO; 33 cases), followed by the government-opposition type of authoritarian-democratic cleavage (AD-GO; 22), the opposition of loyalists to liberals and communists (or the latter two to each other) in the political sphere (AD-2; 17), the opposition of market-oriented individuals to social-paternalists in the socio-economic sphere (SE-SPM; 12) and the opposition of imperialists to anti-imperialists (Syst-IАI; 9).

In 26 cases, the whole political dimensions played the role of predictors, which indicated that confrontation lines in the elite and mass consciousness coincided. Moreover, the first political dimension acted as the predictor in 9 regions, the second acted in 13 regions and the third did in 4. In most cases, these dimensions were dominated by government-opposition variations of socio-economic or authoritarian-democratic cleavage (for more details, see [(Korgunyuk Yu.G. Politicheskiye izmereniya i predmetnye razmezhevaniya: metodika opredeleniya vzaimodeistviya [Political Dimensions and Issue Domain Cleavages: A Methodology for Characterizing Interactions]. – Politiya. 2021. No. 3 (in print). (In Russ.))]).

As for the social connotation, it was found in 155 out of 203 factors of regional dispersion (Table 2); in 6 factors, its statistical significance was 0.1 while in others it was below 0.05. The most common factor of socio-demographic differentiation – on its own or in combination with others – was the level of urbanization (84 cases), followed by demographic characteristics (55), the level of economic activity/independence of population (41), social wellbeing (18), entrepreneurial activity (10), government support of the population or territory (10). Some other, less common factor manifested in 69 cases.

###### Table 2. Regression models of connection between electoral cleavages and factors of social and demographic differentiation
 Region EC R2 Urbanization Demographic characteristics Economic activity (independence) Entrepreneurial activity Social wellbeing Taxpayer support Other Other-2 2016 Republic of Adygea EC-1 0.813 0.784 (0.095) –0.275 (0.102) 0.102 (0.034) Republic of Dagestan EC-1 0.495 0.378 (0.095) –0.630 (0.095) 0.223 (0.093) Republic of Ingushetia EC-2 0.844 –0.696 (0.176) –0.600 (0.176) Republic of Karelia EC-1 0.806 0.523 (0.167) –0.440 (0.167) EC-2 0.322 –0.567 (0.206) EC-3 0.432 0.596 (0.201) –0.466 (0.201) Republic of Mordovia EC-1 0.498 0.706 (0.112) Chechen Republic EC-1 0.402 0.453 (0.188) 0.433 (0.188) Chuvash Republic EC-1 0.841 –0.872 (0.091) –0.264 (0.091) Altai Krai EC-1 0.643 –0.402 (0.112) –0.612 (0.109) –0.269 (0.112) EC-2 0.410 –0.640 (0.136) Krasnoyarsk Krai EC-1 0.783 0.711 (0.056) –0.443 (0.056) 0.135 (0.056) EC-2 0.384 –0.291 (0.094) 0.296 (0.094) 0.219 (0.094) –0.385 (0.094) –0.195 (0.094) EC-3 0.170 –0.258 (0.107) 0.325 (0.107) Perm Krai EC-1 0.846 –0.858 (0.077) 0.201 (0.077) Primorsky Krai EC-1 0.710 0.417 (0.140) –0.524 (0.141) –0.311 (0.137) EC-2 0.641 –0.720 (0.145) –0.349 (0.145) Stavropol Krai EC-2 0.414 0.644 (0.160)

The table is not fully displayed Show table

Only electoral cleavages with the determination coefficients significant at p<0.05 (*p<0.1) are shown.
Electoral cleavages without political interpretation were detected in the following regions:
2019 – Sevastopol (EC 3)

The generalized data on the structure of electoral cleavages in elections to regional assemblies and trends in structural changes are described in Table 3, which is based on the type of structure (column 3), that is starting with the most dense ("complete") and ending with the least dense ("incomplete" ECs with one characteristic) structure.

###### Table 3. Types of structures of electoral cleavages in the legislative assembly elections (2016–2020)
 Year Number of parties Structure type Structure Region Trend 2016 5 Complete 11 Kursk Oblast 1 2016 9 Complete 11 Lipetsk Oblast 0 2016 5 Complete 11 Republic of Mordovia 1 2016 4 Complete 11 Oryol Oblast 1 2016 5 Complete 11 Republic of Adygea 1 2016 5 Complete 11_11 Kamchatka Krai 1 2016 5 Complete 11_11 Tambov Oblast 1 2017 6 Complete 11_11 Saratov Oblast 0 2018 8 Complete 11_11 Ivanovo Oblast 1 2018 5 Complete 11_11 Kemerovo Oblast -1 2018 7 Complete 11_11 Rostov Oblast 1 2020 8 Complete 11_11 Voronezh Oblast 1 2016 7 Complete 11_11_11 Saint Petersburg 1 2020 10 Complete 11_11_11 Ryazan Oblast 1 2016 10 Rich 11_11_01 Krasnoyarsk Krai 0 2018 7 Rich 11_00_11 Irkutsk Oblast -1 2020 11 Rich 01_11_11_00 Kaluga Oblast 1 2016 13 1st complete + ideol + soc 11_10_01 Moscow Oblast -1 2017 6 1st complete + ideol + soc 11_10_01 Udmurt Republic 0

The table is not fully displayed Show table

The fourth column (structure) indicates the configuration of electoral cleavages in each region: 1 indicates a feature, 0 indicates no feature; the figures are broken down into pairs – the first indicates the presence / absence of a political interpretation, the second indicates the presence / absence of a social connotation. For example, the "11_01_00" configuration indicates that the elections revealed three factors of territorial dispersion, the first of which is a "complete" electoral cleavage, the second only has a social connotation, and the third has neither political content nor social connotation.

The sixth column demonstrates the structural change trends for electoral cleavages. "1" indicates consolidation, "-1" indicates erosion, "0" indicates no visible changes.

The table shows that the most common type of structure of electoral cleavage was one in which only the first EC is "complete", and the rest have a social connotation only (30 cases). Next in terms of prevalence are the "complete" (14) and "incomplete" structure with cleavages having only a social connotation (11). The rest of the types are represented by isolated cases.

As for the change in the structure of electoral cleavages, in 51 regions out of 83 indicated an erosion trend, 19 indicated a consolidation trend and 13 did not indicate any serious transformations. If we break down all cases by year, we find that the erosion trend has always prevailed, but with varying intensity: it was the most intense in 2018 (12 regions out of 15) and the least intense in 2020 (6 out of 11). The consolidation trend was most pronounced in 2016 (12 out of 39) and in 2020 (4 out of 11).  In 2017 and 2019, there were no such cases whatsoever, but it was during these years that every third region retained the same cleavage structure.

Let us take a closer look at each cleavage structure.

A distinctive feature of the structure that includes complete cleavages only is that consolidation took place in 11 cases out of 14, and erosion took place only in Kemerovo Oblast, and even there it was rather arbitrary. In Kemerovo Oblast, only one EC manifested in the Duma elections, while two did in the 2018 regional elections, and the fragmentation of a single EC was accompanied by an increase in the effective range, politicization and socialization coefficients of both electoral cleavages; in many respects this can be considered a consequence of the increased political competition after the departure of Aman Tuleyev, the region's governor of over 20 years.

Five regions of this group indicated only one electoral cleavage. Four parties were admitted to the elections Oryol Oblast, five in the Republic of Adygea, the Republic Mordovia and Kursk Oblast as well as nine in Lipetsk Oblast (it is obvious that these regions were part of the Red Belt in the 1990s). All elections were held in 2016, at the same time as the Duma elections.

Lipetsk Oblast was the only region to indicate the cleavage structure in regional elections as nearly the same as in the federal election. In other regions, the Duma election indicated more territorial dispersion factors, and one of said factors was typically an incomplete electoral cleavage. In all cases except for Mordovia, the effective range, politicization and socialization coefficients in regional elections were significantly higher than in the federal election. The government vs. the opposition in the socio-economic sphere (SE-GO) was the most common political confrontation of the only electoral cleavage everywhere. This confrontation sometimes combined with other confrontations (liberals vs. communists in the socio-economic sphere, imperialists vs. anti-imperialists) and the prevailing socio-economic background, like the level of urbanization, which also sometimes combined with other factors of inter-territorial differentiation.

Seven regions indicates two complete cleavages. Five parties were admitted to the elections in Kamchatka Krai, Kemerovo and Tambov Oblasts, six in Saratov Oblast, seven in Rostov Oblast, eight in Ivanovo and Voronezh Oblasts. Most of these regions (except Kamchatka Krai and Ivanovo Oblast) used to be part of the Red Belt as well.

The most common variation of political content of the first electoral cleavage was the government vs. the opposition confrontation in the political sphere (AD-GO), followed by the same confrontation in the socio-economic sphere (SE-GO). Urbanization level was the prevalent one among social predictors. As for the second EC, its political identity was determined by the liberals vs. communists confrontation in the socio-economic sphere (SE-LiC) in Tambov Oblast only. It should be noted that the federal elections are characterized precisely by the fact that the first EC has authoritarian-democratic connotations, the second is characterized by the liberals vs. communists confrontation in the socio-economic sphere, and that urbanization level is the main socio-demographic predictor for both of them [11].

We may therefore conclude that the cleavage between market supporters and social-paternalists in modern Russia stems from the differences between large cities and small towns rather than the differences between urban and rural areas. When identifying electoral cleavage structure at the federal level, federal subjects are used as territorial units, and the level of their urbanization is determined primarily by the development of large cities. However, if municipal okrugs (districts) are used as territorial units, then the differences between rural and urban areas typically form an electoral cleavage marked by the government vs. the opposition confrontation only.

Finally, two regions – Saint Petersburg and Ryazan Oblast – indicated three complete electoral cleavages at once. Only Ryazan Oblast used to be a part of the Red Belt; Saint Petersburg, on the contrary, is rightly considered as the most liberal region of the country, and this difference can be easily seen in the political connotation of electoral cleavages. In Saint Petersburg, the first EC is associated with the imperialists vs. anti-imperialists confrontation (Syst-IAI), the second – with the liberals vs. communists confrontation in the socio-economic sphere (SE-LiC) and the liberals vs. loyalists in the political sphere (AD-LiLo), the third – only with the liberals vs. loyalists confrontation (AD-LiLo). In Ryazan Oblast, where Yabloko was not allowed to participate in the elections, the first electoral cleavage had a complex political nature – there were signs of confrontation between Soviet traditionalists and progressives (Syst-STP), loyalists and the opposition in the political sphere (AD-LoOp), the government and the opposition in the socio-economic sphere (SE-GO). The second and the third ECs centered around the confrontation between loyalists and the opposition in the political sphere (AD-LoOp). In Saint Petersburg, the electoral cleavage, among other things, had more solid effective range, politicization and socialization coefficients. This allows us to assume that the "complete" structure of the EC in Saint Petersburg is explained by the high level of political awareness among voters there, while the complete structure in Ryazan Oblast is likely incidental.

The group with a "rich" structure of electoral cleavages is small and includes only three regions: Krasnoyarsk Krai (10 participating parties), Irkutsk Oblast (7) and Kaluga Oblast (11). The first two regions indicated three electoral cleavages while the last one indicated four. However, only two ECs were complete everywhere: the first and the second in Krasnoyarsk Krai, the first and the third in Irkutsk Oblast, and the second and the third in Kaluga Oblast. Moreover, each of the regions showed different trends in the evolving EC structure: consolidation in Kaluga Oblast, erosion in Irkutsk Oblast, and stability in Krasnoyarsk Krai.

The most numerous group turned out to be one, where only the first electoral cleavage (40) was complete. In the most common structural pattern of the cleavages, all other ECs only had a social connotation without a clear political one (30). The number of participating parties in this case was quite varied: from 4 in Chukotka Autonomous Okrug to 13 in Moscow Oblast.

The most common variation of political connotation of the first EC was the government vs. the opposition confrontation, either in socio-economic (SE-GO) or political sphere (AD-GO), and the most widespread social connotation was the level of urbanization. As for trends, only 3 cases out of 40 indicated consolidation: Jewish Autonomous Oblast and Chukotka Autonomous Okrug (both in 2016) as well as Magadan Oblast (2020). Eight more regions did not indicate any change, and 29 (almost 75%) indicated erosion.  In this group, only Bryansk, Penza and Ulyanovsk Oblasts used to be part of the Red Belt.

There were nine regions in the group where the only complete electoral cleavage never placed first in the hierarchy; the number of participating parties varied from 5 in Stavropol Krai and Novgorod Oblast to 10 in Kostroma Oblast. The complete electoral cleavage placed second in nearly every region, and it placed third in Kostroma Oblast only. The more common variation of political connotation of this cleavage was the government vs. the opposition confrontation, either in socio-economic (SE-GO) or political sphere (AD-GO). The Komi Republic (2020) was the only region where the cleavage between Soviet traditionalists and progressives (Syst-STP) functioned in this capacity, while in the Republic of Buryatia (2018) it was communists vs. loyalists cleavage in the political sphere (AD-LoC). Like in the rest of the cases, urbanization level was the most common social connotation of the cleavage. Erosion was the most common trend in structural transformation of electoral cleavages (five regions out of nine), although one third of the cases (Stavropol Krai, Novgorod Oblast, Tomsk Oblast – all in 2016) indicated the opposite trend.

Finally, 17 regions indicated only "incomplete" electoral cleavages, and five Russian regions indicated that part of the EC had a political interpretation only. Some regions only indicated a social connotation; one (Sevastopol) indicated that one of the ECs (the second) only had a political interpretation. The remaining 11 cases indicated that the existing factors of territorial dispersion only had a social connotation. All regions of this group indicated a trend towards erosion, except for the Kabardino-Balkarian Republic, where the structure of electoral cleavages have not changed as compared to the 2016 Duma election.

It is interesting that the group of regions where electoral cleavages only had a social connotation included regions with both shaky electoral reputations (Chechnya, Dagestan, Ingushetia) and fairly high competitive levels (Khakassia, Karelia, Altai Republic, Amur Oblast). In both places, the leading factors of territorial dispersion indicated a confrontation between United Russia and the majority of other election participants, but even relatively high modules of correlation coefficients were nullified by statistical significance that exceeded the maximum allowed 0.1. We may assume that in the North Caucasus, the results of certain parties were simply "adjusted", while in "competitive" regions, voters cast their votes based on personal preferences for certain regional leaders rather than political preferences.

## Conclusion

After comparing electoral cleavages in the 2016 Duma election and in the 2016–2020 regional legislative assembly elections, we may report that erosion was the predominant trend in structural changes of cleavages, despite the fact that in regional elections, there were typically fewer participating parties than in the Duma election.

This also applies to 2016, when elections to regional legislative assemblies were held at the same time as the Duma election. Even then, the structure of electoral cleavages in regional elections was prone to erosion in more than half of the regions. At the same time, nearly one third of the cases indicated consolidation of the EC structure. In other words, transformation trends were contradictory.

In the three years that followed (2017–2019), the trend of eroded electoral cleavages became more pronounced: only two out of 33 regions indicated consolidation, while seven more managed to maintain a status quo. To some extent, this may be explained by denying election registration to liberal candidates (parties like Yabloko and PARNAS), whose absence in the elections eliminated the external policy dimension, which is essential in ensuring political polarization.

In 2020, however, more than a third of the regions indicated a trend for consolidation of electoral cleavage structure, despite the fact that neither Yabloko nor PARNAS were allowed to participate in the elections there. However, participants included the newly formed Direct Democracy Party, as well as the New People and For Truth parties, which demonstrated fairly clear ideological leanings: the former two belonged to the liberal side of the spectrum, while the latter represented social patriotism.

This partly confirms the assumption that the administrative regulation of the composition of electoral participants under proportional representation system can blur or, on the contrary, clarify the political picture for the voter. Eliminating one of the opposition poles represented by Yabloko and PARNAS typically blurs the picture (although this rule does not apply to the Red Belt regions), while the new liberal contenders for seats in (Direct Democracy Party, New People) makes it clearer.

However, it seems that the situation is even more complicated. It would be a logical assumption that in different regions of the country, mass political consciousness stands at different development levels. For example, in Chechnya, Dagestan and Ingushetia, the mass consciousness does not yet think in political categories, but in clan affiliations instead. In Karachay-Cherkessia and Kabardino-Balkaria, the mass consciousness is ready to comprehend the government-communist opposition as the vein of political life; in Adygea, this comprehension even has a social connotation. The confrontation between the "pro-market" government and the anti-market left and nationalist opposition takes the most complete forms in the former Red Belt regions (in the Central Black Earth Region before all else) while becoming more and more blurred in others. For example, in the relatively competitive Republic of Karelia, people cease to perceive the previous political landscape, but do not see any alternative yet. On the other hand, in Saint Petersburg, with its rather politically savvy electorate, political landscape has entered a new stage of development. In this case, we can observe a new structure of electoral cleavages – one that has clear political leanings and a pronounced social connotation.

The evolution of mass political consciousness is nonlinear, and from this point of view, different Russian regions often develop in opposition to each other. Whether or not this true, and how the indicated trends will develop, we will see in September 2021, when the next elections to the State Duma and regional legislative assemblies take place.

#### References

1. Akhremenko A.S. Struktura elektoralnogo prostranstva [Electoral Space Structure]. Moscow: Sotsialno-politicheskaya mysl. 2007. 320 p. (In Russ.)
2. Bornschier S. Cleavage Politics and the Populist Right: The New Cultural Conflict in Western Europe. Philadelphia: Temple University Press, 2010. 345 p.
3. Franklin M. The decline of cleavage politics. – Electoral change: Responses to evolving social and attitudinal structures in Western countries / Eds. Franklin M., Mackie T., Valen H. Cambridge, UK: Cambridge University Press, 1992. P. 381–402.
4. 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.
5. 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.
6. 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.
7. Inglehart R. Cultural shift in advanced industrial society. Princeton, NJ: Princeton University Press, 1990. 484 p.
8. Korgunyuk Yu. Cleavage Theory and Elections in Post-Soviet Russia. – Perspectives on European Politics and Society. 2014. Vol. 15. No. 4. P. 401–415.
9. Korgunyuk Yu.G. Kontseptsiya razmezhevanii i faktorny analiz [Cleavage Theory and Factor Analysis]. – Politiya. 2013. No. 3. P. 31–51. (In Russ.).
10. Korgunyuk Yu.G. Kontseptsiya razmezhevanii i teoriya problemnykh izmerenii: tochki peresecheniya [Cleavage Theory and Problematic Dimensions Theory: Points of Intersection]. – Polis. 2019. No. 6. P. 95–112. DOI: 10.17976/jpps/2019.06.08. (In Russ.) - https://doi.org/10.17976/jpps/2019.06.08
11. 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/
12. 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/
13. Korgunyuk Yu.G. Politicheskiye izmereniya i predmetnye razmezhevaniya: metodika opredeleniya vzaimodeistviya [Political Dimensions and Subject Cleavages: A Methodology for Characterizing Interactions]. – Politiya. 2021. No. 3 (in print). (In Russ.)
14. Korgunyuk Yu.G. Vybory po proportsionalnoi sisteme kak massovy opros obshchestvennogo mneniya [Elections under Proportional Representation System as a Public Opinion Poll]. – Politicheskaya nauka. 2017. No. 1. P. 90–119. (In Russ.)
15. Korgunyuk Yu. Issue Dimensions and Cleavages: How the Russian Experience Helps to Look for Cross-points // Russian Politics. 2020. No. 5. P. 206–235. DOI: 10.30965/24518921-00502004. - https://doi.org/10.30965/24518921-00502004
16. Kriesi H., Grande E., Dolezal M., Helbling M., Hoglinger D., Hutter S., Wuest B. Political Conflict in Western Europe. Cambridge: Cambridge University Press, 2012. 349 p.
17. 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.
18. New politics in Western Europe: The rise and success of Green Parties and alternative lists / Ed. Müller-Rommel F. Boulder, CO: Westview, 1989. 230 p.
19. Sata R., Karolewski I.P. Caesarean politics in Hungary and Poland. – East European Politics. 2020. Vol. 36. No. 2. P. 206–225.
20. 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 27.04.2021). - http://www.polsci.ucsb.edu/faculty/hstoll/research/thesis.pdf
21. 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. 575–95.
22. Torcal M., Mainwaring S. The Political Recrafting of Social Bases of Party Competition: Chile, 1973–95. – British Journal of Political Science. 2002. Vol. 33. No. 1. P. 55–84.