THE NATIONAL COUNCIL FOR EURASIAN AND EAST EUROPEAN RESEARCH TITLE VIII PROGRAM

Transcription

1 TITLE: URBAN-RURAL VOTING DIFFERENCES IN RUSSIAN ELECTIONS, : A RAYON LEVEL ANALYSIS AUTHORS: RALPH S. CLEM and PETER T. CRAUMER Florida International University THE NATIONAL COUNCIL FOR EURASIAN AND EAST EUROPEAN RESEARCH TITLE VIII PROGRAM 1755 Massachusetts Avenue, N.W. Washington, D.C

2 LEGAL NOTICE The Government of the District of Columbia has certified an amendment of the Articles of Incorporation of the National Council for Soviet and East European Research changing the name of the Corporation to THE NATIONAL COUNCIL FOR EURASIAN AND EAST EUROPEAN RESEARCH, effective on June 9, Grants, contracts and all other legal engagements of and with the Corporation made under its former name are unaffected and remain in force unless/until modified in writing by the parties thereto. PROJECT INFORMATION:' CONTRACTOR: PRINCIPAL INVESTIGATOR: Florida International University Ralph Clem COUNCIL CONTRACT NUMBER: DATE: January 20, 1998 COPYRIGHT INFORM A TION Individual researchers retain the copyright on their work products derived from research funded by contract with the National Council for Eurasian and East European Research. However, the Council and the United States Government have the right to duplicate and disseminate, in written and electronic form, this Report submitted to the Council under this Contract, as follows: Such dissemination may be made by the Council solely (a) for its own internal use, and (b) to the United States Government (1) for its own internal use; (2) for further dissemination to domestic, international and foreign governments, entities and individuals to serve official United States Government purposes; and (3) for dissemination in accordance with the Freedom of Information Act or other law or policy of the United States Government granting the public rights of access to documents held by the United States Government. Neither the Council, nor the United States Government, nor any recipient of this Report by reason of such dissemination, may use this Report for commercial sale. 1 The work leading to this report was supported in part by contraa funds provided by the National Council for Eurasian and East European Research, made available by the U. S. Department of State under Title VIII (the Soviet-Eastern European Research and Training Act of 1983, as amended). The analysis and interpretations contained in the report are those of the author(s).

3 CONTENTS Executive Summary i Introduction 1 Regional Patterns of Rural-Urban Voting 2 Urban-Rural Voting Differences at the Rayon Level 4 Urban-Rural Voting Differences: Evidence from the 1995 and 1996 Elections 5 Party Preferences: Presidential Elections: Voter Turnout 10 Differences of Voting Behavior Among Rural Areas 11 Regional Contextual Effects: Socioeconomic Correlates 11 Summary and Conclusion 14 References 15 Tables 1 through 7 17

4 Executive Summary Since the collapse of the Soviet Union, the Russian electorate has gone to the polls nation-wide five times. Taken together, these events provide a body of evidence from which some very interesting spatial patterns of electoral behavior may be discerned, In this paper, our focus is on the regional aspect of voting differences between urban and rural areas of Russia, a divide which separates the population along political lines perhaps more distinctly than any other single factor. We explore possible explanations behind the urban-rural electoral split, adducing other aggregate socioeconomic and demographic characteristics of the population which co-vary with the degree of urbanization (or its reverse, "ruralness"). To measure the regional and urban-rural aspect of voting behaviorin post-soviet Russia, we utilize here results of the secondnational Duma election in December, 1995 and the presidential elections of June and July, Specific measures of voting behavior employed here are. preference for different political parties in the Duma election; and choices made among candidates in the June, 1996 presidential primary and in the runoff election between Yel'tsin and Gennadiy Zyuganov in July, 1996, We also examine regional and urban-rural variations in voter turnout. Disaggregating the 1995 voting data in our nine-oblast set, we found the same pattern of urban-rural and degree of ruralness differences in party preference as was manifested in the election. Specifically, as in 1993, the vote for reform bloc of parties was consistently higher in urban rayons. and fell in rural rayons along the degree of ruralness slope. Also replicating the 1995 pattern, left parties did considerably better in rural areas than in cities. Finally, in 1995 the parties of the right showed much less variation across the urban-rural divide and by degree of ruralness, just as was the case in the 1993 contest. There were, however, some important changes in the magnitude of the vote for the different party groupings, The percentage of the vote for the reform parties dropped significantly and that for the left increased dramatically among the units in our data set. This mirrors the nation-wide result, wherein the KPRF and other left parties increased their share of the proportional representation vote by about 12 percentage points, the reform parties dropped by almost four percentage points, and the right fell by approximately three percentage points. Even more significantly, the largest declines in percentage of the vote between 1993 and 1995 for the reform parties was in the urban areas- in 1993 these parties accrued 38.6 percent of the vote in cities within our nine-oblast sample, but by 1995 that share had fallen to 26.4 percent, suggesting a dangerous erosion of the traditional base of support for the reformers, On the other hand, the inroads made by parties of the left in regions of Russia which had favored the reform parties in prior elections and referenda is also reflected in our rayon data; the share of the vote going to the left in the cities of our sample jumped from just 11.3 percent in 1993 to 30.7 percent in Voting for left parties in rural rayons also increased appreciably between 1993 and 1995, contributing still further to their overall strong showing.

5 Results of the national party list vote disaggregated by size of city for the December 1995 Duma election generally conform to the pattern of party preference established in earlier electoral events. As in our study of the 1993 Duma election, preference for the reform parties is greatest in the largest cities (the ten cities of 250,000 people or more), falls with decreasing size of urban centers, then unexpectedly (i.e., not previously seen) rises in the smallest towns. On the other hand, as was true in prior elections and referenda, support for the left is relatively weaker in the largest cities, then rises as city/town size decreases. Here again, however, the result in the smallest towns is anomalous, as strength among the left parties dropped in places with fewer than inhabitants. Parties of the right enjoy comparatively constant upport across city size, excepting again in the smallest towns, where their percentage increased. As was true with the rayons, voter preference for the party blocs changed markedly between 1993 and 1995 by size of city. Thus, in the largest cities (those with 250,000 inhabitants and over) the percentage of the vote for the reform parties dropped by over ten percentage points (43.0 to 32.6 percent) while that for parties of the left soared from about 11 percent to over 28 percent. In medium-sized cities (50,000 to 250,000), the shift from reform to left was even more stunning, in those cities, the left actually outpolled the reformers in Disaggregating the June 16 presidential election figures along urban-rural lines produces exactly the result one would have expected given the experience of prior elections. Yel'tsin was the top vote-getter in the cities, but his strength ebbed in rural areas. Contrariwise. Zyuganov lagged Yel'tsin in the cities, but outperformed him in the more rural rayons. Yavlnskiy was relatively more popular in cities, just as his eponymous party Yabloko has been, while, as before, Zhirinovskjy's support was almost evenly distributed (albeit at a low level) right across the urban-rural continuum. Voting outcomes by city size largely confirm hypothesized trends. Yel'tsin dominated the largest cities, while Zyuganov managed only a quarter of the vote. Yavlinskiy also did best in the largest cities, but Lebed and Zhirinovskiy were most popular in medium-sized cities. The most interesting result among the top five finishers is Lebed's appeal to city voters; it has been suggested that Lebed drew support mainly from an amorphous body of uncommitted voters who were reluctant to support "pure" reform parties and candidates, but also were alienated from the nationalist and left parties. Given that rural dwellers have traditionally voted left or right, that constituency would have held less promise for Lebed, but there was obviously a previously untapped well of voters in cities for which he was an attractive candidate. Among our sample oblasts, results of the Yel'tsin-Zyuganov runoff matched the national totals almost exactly: 54.4 percent for Yel'tsin and 40,4 percent for Zyuganov. Disaggregated by rayon, the vote tally also conformed to the expected pattern.- that is, Yel'tsin did relatively better in the more urbanized rayons and Zyuganov in the more rural areas. In fact, Zyuganov actually bested Yel'tsin in all but the most urbanized and, anomalously, the most rural rayons; in the case of the

6 most rural rayons (those over 75 percent rural), the weight of the populous Sverdlovsk Oblast biases the overall result in Yel'tsin's favor. Considering the runoff results by size of city, the dominance of Yel'tsin in urban areas is clear, especially in the largest cities where his margin of victory was almost 25 percentage points. Our results also provide some clues as to the question of vote shifting from the June 16 primary to the July3 runoff, a subject of considerable interest to students of Russian politics. We correlated the percentage point change in the vote for Yel'tsin by unit between the two rounds with the percentage of the vote for Lebed in round one, and found a coefficient of.769, indicating that units in which Lebed did well on June 16 evinced a major increase in the vote for Yel'tsin on July 3. suggesting a shift in votes from the general to the president. The coefficient for change in the vote for Zyuganov and voting for Lebed is considerably smaller (.361 ), indicating that the former picked up relatively fewer votes from the latter. This same pattern of movement from Lebed to Yel'tsin was observed by Myagkov, Ordeshook, and Sobyanin in their exhaustive study of rayon-level voting trends. Our data also show some shifting of votes from Yavlinskiy to Yeltsin (a correlation of.651). While there is no doubt that urban-rural residence is a major consideration in the analysis of voting behavior in Russia, interregional differences in electoral choices suggest that a host of other influences shape the manner in which people vote. Most importantly, the regional context, especially the economic situation as translated through income, appears to play a powerful role, as variations in wealth across rural areas engender different responses to issues, candidates and parties. Other, local concerns, different degrees of sophistication of party organization, the talents and reputations of individual candidates, and other difficult-to-quantify factors likewise enter into the calculus of voting behavior. Finally, several structural variables other than urban/rural including those such as education and occupation which are inextricably bound up with urbanization, must be taken into account when assessing electoral results. in

7 URBAN-RURAL VOTING DIFFERENCES IN RUSSIAN ELECTIONS, : A RAYON-LEVEL ANALYSIS RALPH S. CLEM AND PETER R. CRAUMER 2 Since the collapse of the Soviet Union, the Russian electorate has gone to the polls nation-wide five times. Taken together, these events provide a body of evidence from which some very interesting spatial patterns of electoral behavior may be discerned. In this paper, our focus will be on the regional aspect of voting differences between urban and rural areas of Russia, a divide which separates the population along political lines perhaps more distinctly than any other single factor. We will explore possible explanations behind the urban-rural electoral split, adducing other aggregate socioeconomic and demographic characteristics of the population which co-vary with the degree of urbanization (or its reverse, "ruralness"). To measure the regional and urban-rural aspect of voting behavior in post-soviet Russia, we utilize here results of the second national Duma election in December and the presidential elections of June and July, Specific measures of voting behavior employed here are: preference for different political parties in the Duma election; and choices made among candidates in the June, 1996 presidential primary and in the runoff election between Yel'tsin and Gennadiy Zyuganov in July, We will also examine regional and urban-rural variations in voter turnout. Details of the methodology and data involved will be found in our previously published work (Clem and Craumer, 1993; 1995a; 1995b; 1995c; 1996). As will be discussed below, a clear and remarkably stable electoral geography of Russia has emerged in the post-soviet period. 3 Using aggregate data at different geographical levels of analysis. one finds a pronounced regionalization of voting behavior across Russia. This is entirely in keeping with the experience of virtually all countries, certainly all large countries, wherein different regional "political ecologies" provide the context which shapes the manner in which people behave politically (Agnew, 1987; Brustein, 1990; O'Loughlin, Flint, and Anselin, 1994). From the point of view of the political geographer, the "...contexts within which voters [make] their decisions [are] multidimensional, with multilayered composites of structural (socioeconomic) and spatial influences" (O'Loughlin, Flint, and Anselin, 1994: 372). Thus, in a large and diverse country such as Russia there is little credence attached to the notion of a "national" electorate. Rather, socioeconomic 2 Professor and Associate Professor, respectively, Department of International Relations, Florida International University, Miami, FL The authors wish to acknowledge funding support from the National Council for Eurasian and East European Research, which however is not responsible for the contents or findings of this paper. We are also grateful to Andrei Berezkin for his advice and Judith Rasoletti for her assistance. 3 In fact, an emerging electoral geography of Russia was evidenced as far back as the elections to the Union-wide Congress of People's Deputies in 1989, to the RSFSR Congress of People's Deputies in 1990, and the Russian Federation presidential election in See: Berezkin, et al, 1989, and Kolosov, Petrov, and Smirnyagin,

8 variables must be taken together with place-specific, or contextual, effects to even approach a satisfactory understanding of variations in voting across regions. Several reasons have been advanced to explain why structural variables alone so often fail to account for the regional diversity in electoral behavior. Such considerations as "...the differential impact of party organization, extended family and social networks, the appeal and vigor of candidates beyond the party attraction, local campaign effects, party competition, and local political culture" may sway voters across socioeconomic lines (O'Loughlin, Flint, and Anselin, 1994: 352). National-level party platforms may resonate very differently from one region to the next, depending on local economic conditions, agricultural and industrial vested interests, import-export-based economies, sheer location vis-a-vis neighboring states, international trade linkages, and/or resource exploitation. National economic trends typically impact regions in large countries very differently, with there almost always being clear winning and losing regions from both overall growth or decline. Having said this, urban-rural voting differences in Russia are so consistent as to make this the best starting point for disaggregating the vote; indeed, as will be seen, voting behavior is tied not only to the urban-rural dichotomy, but to degrees of the "urbanness" or "ruralness" of spatial units. Surveys of the Russian population reveal pronounced variations in attitudes toward reform or antireform issues and/or parties between urban and rural residents and by size of place (Hough, 1994; Hough, Davidheiser, and Lehmann, 1996). Indeed, Lehmann has shown that attitudes toward marketization vary not only according to urban-rural residence, but also according to one's birth in the city or countryside and the age at which one had moved to the city (Hough, Davidheiser, and Lehmann, 1996: 6). After all, Russia is a country which has undergone very rapid urbanization in this century. perhaps the most rapid in history (Lewis and Rowland, 1979: ). Along with this dramatic shift of the Russian population to urban centers, there has occurred a widening of the gap between the city and the countryside across a broad range of social attributes; with regard to such social traits as occupation and levels of education, cities have become fundamentally distinct places compared to rural areas. Likewise, the demographic structure of the urban population tends to be much younger, and fertility is typically much lower in the cities. Yet, as will be seen below, urbanization as a single structural variable leaves large residuals in terms of accounting for regional voting patterns. Regional Patterns of Urban-Rural Voting Beginning with the referendum of April, 1993, major inter-regional variations in voting behavior have been evident within Russia, variations which are tied to a significant degree to the extent to which regions are urbanized (Clem and Craumer, 1993: ; Petrov, 1996; Slider, Gimpel'son, and Chugrov, 1994: 730). Specifically, support for Yel'tsin and his economic policies

9 was strongest in the largest urban centers (Moscow and St. Petersburg) and areas of the country with comparatively high levels of urbanization (principally those in the Urals, the Center, the North and Northwest, and individual units such as Kirov and Nizhegorod oblasts. On the other hand, units with a higher percentage of rural population, such as those along the western border, in the Central Chernozem, across the Volga and along the southern frontier with Kazakhstan, tended to be less supportive of Yel'tsin and economic reform. This same broad pattern manifested itself again in the December, 1993 parliamentary election and constitutional plebiscite (Clem and Craumer, 1995a). Highly urbanized regions registered greater support for the pro-reform political parties on the national party list vote, whereas the comparatively more rural units voted in much larger numbers for the anti-reform parties [i.e., for the Agrarian- Communist coalition (Agrarian Party of Russia, or APR, and Communist Party of the Russian Federation, KPRF) or for the Liberal Democratic Party of Russia (LDPR) headed by Vladimir Zhirinovskiy]. 4 The vote on the Yel'tsin constitution in December likewise showed a pronounced urban-rural split, with urbanized regions voting more in favor and rural regions more against. To illustrate the consistency of the voting patterns which have developed in the recent Russian elections, there was a strong statistical relationship (.748, significant at the.01 level) among the 87 units of the Russian Federation which participated in both the April, 1993 referendum and December, 1993 constitutional plebiscite in terms of the support shown to Yel'tsin in the former and to his proposed constitution in the latter (Clem and Craumer, 1995a). It has become popular to characterize this geographical pattern as the "North-South Divide", with the 55th North parallel usually used as the boundary between regions supportive of reform (to the north) and those opposed (to the south) of that line (Orttung and Parrish, 1996; Slider, Gimpel'son, and Chugrov. 1994: ). Although this is a convenient generalization, the operative variables are, in the first instance, the level of urbanization between the more industrialized north and the more agricultural south; secondly, the other socio-structural and economic differences largely subsumed by urbanization (i.e.. income, education, occupation, and age); and, thirdly, local or contextual effects. Despite the much-publicized gains made by parties of the left on the proportional representation balloting in the December, 1995 parliamentary contest, once again the same general outlines of regional political affiliations emerged. The statistical relationship between urbanization and party preference (as measured by the proportional representation vote) degraded slightly between 1993 and 1995, mainly because the left parties gained some ground in the more highly urbanized units which had previously evinced higher levels of support for the reform parties (Clem and * Party preference in this paper refers only to the national party list, or proportional representation, component of the Duma elections. The other half of the 450-seat Duma is elected from single-member districts by plurality vote. The political affiliation of candidates in the single-member district elections has not always been clear, and thus it is difficult to use these results to determine party preference.

10 Craumer, 1995c: ). Finally, the two 1996 presidential contests produced a clear north-south split, with the north voting more heavily for Yel'tsin and the south for Zyuganov (Clem and Craumer, 1996; TsIK, 1996). However, the urban variable, which had proven so powerful in the past, weakened considerably as a correlate of voting choices (Clem and Craumer, 1996: 362). Voter Turnout Although the subject has not received much attention from scholars, there have been major differences among the regions of Russia in voter turnout (Clem and Craumer, forthcoming). For example, in April 1993, turnout ranged from a high of 78.8 percent (Ryazan Oblast) to a low of 54.3 percent (in the Khanty-Mansiy Autonomous Okrug); the national turnout rate was 64.5 percent (Clem and Craumer, 1993: ). Like voter choices on the referendum, plebiscite, and party preference, voter turnout on the regional scale has related to levels of urbanization: relatively highly urbanized areas typically manifest a lower turnout rate, whereas a much higher percentage of voters have turned out in the more rural units. There is also a high to moderate consistency in voter turnout among regions over time: the correlation coefficient for turnout by unit from April 1993 to December 1993 was.813, and from December 1993 to December 1995,.674, both of which are significant at the.01 level (Clem and Craumer. 1995a: 80; 1995c: 608). Urban-Rural Voting Differences at the Rayon Level Although the broad regional pattern of electoral choices at the oblast and equivalent unit level described above suggests some relationship between voting behavior and urbanization, clearly one cannot draw any definitive conclusions based on simple correlations at such a gross geographical scale of analysis. Accordingly, in a previous study we disaggregated voting data for the April, 1993 referendum and the December, 1993 Duma election and constitutional plebiscite for a nine-oblast set down to the next lowest level, the rayon scale, to examine more precisely how voting behavior related to the degree to which these much smaller units are urbanized (Clem and Craumer, 1995b). Rayons (rayoni) are the smallest political-administrative units in the Russian Federation for which election results and social and economic data are given; although they vary widely in size and population, they might be thought of as roughly equivalent to counties in the United States. Rayons are of two general types: relatively large areal units and sub-divisions of urban areas (gorodskie rayoni), which, together with a number of cities which have an "independent" administrative status, comprise the some 2,700 territorial electoral commissions (territorial'nyye izbiratel'nyye komissii) of Russia. 5 Our nine sample oblasts (Figure 1) were selected to represent regions of the country with 3 The territorial electoral commissions are aggregated within member units of the Russian Federation to form the single-member districts of the Duma.

11 different political, social and economic characteristics; within these nine oblasts there were 303 rayons which formed the data set for the 1993 study. 6 The 1996 population of the nine-oblast set was 17.3 million, representing about 12 percent of the total population of Russia; oblast populations within the set range from 648,000 (Sakhalin) to 4, (Sverdlovsk) (Goskomstat Rossii, 1996a: ). In examining the rayon level data for the two electoral events of 1993, it became apparent that appreciable differences existed within oblasts and other units of that size as regards the various choices made by the electorate; in fact; in some cases the intra-oblast differences were as great as those among oblasts (Clem and Craumer: 1995b: ). The rayon-level analysis of the 1993 data provided some interesting insights into the relationship between the setting in which one resides and the manner in which one votes. For the 1993 data, a strong urban-rural divide is noted for every political choice variable except support for the LDPR. Thus, the percentage of those voting for the APR/KPRF coalition was over twice as high in rural rayons than in cities, whereas support for Yel'tsin on the April referendum, support for his constitution, and support for the reform parties was higher in urban than in rural rayons. Moreover, these voting tendencies become stronger as one moves from lowest to highest degrees of ruralness. Thus, only 43 percent of those in the most rural rayons (those over 75% rural) approved of Yel'tsin on the April, 1993 referendum, compared to 66.6 percent in cities. The 1993 figures also point to large differences in voter turnout along the urban-rural spectrum: turnout tended to be much higher in the most rural rayons (Clem and Craumer, 1995b: ). With the detailed data in the 1993 set, we were also able to disaggregate the urban category by size of city. The results, which we also presented in our earlier study, suggest that "...the political gradient runs straight through the urban component in almost linear fashion" (Clem and Craumer: 1995b: 470). In other words, support for Yel'tsin and the reform parties is highest in the largest cities (i.e., those over 250,000 population), and declines through the medium to small cities and merges into the trend that runs through the rural rayons by degree of ruralness. Turnout likewise conforms to the size of places, with the lowest figures in the largest cities and increasing with decreasing size of cities. Urban-Rural Voting Differences: Evidence from the 1995 and 1996 Elections We have recently completed a rayon-level analysis of differences in voter preference and turnout for the Russian Duma election of December 1995 and the presidential races of 1996, and present the results here for the first time. We employ the same nine-oblast set; however, changes in 6 The oblasts in the set are: Murmansk, Novgorod, Bryansk, Kirov, Kursk, Saratov, Sverdlovsk, Kemerovo, and Sakhalin. All but two (the North Caucasus and East Siberia) of the macro-regions of Russia are represented. We avoided selecting any non-russian republics to minimize the impact of the non-russian ethnic factor.

12 the alignment of rayons reduced the number of units in the analysis to 292 for 1995 and 300 for Changes in the number or composition of units for 1995 and 1996 is attributed to a drop in the number of cities; there were 70 cities in the 1993 set, 61 in the 1995 set, and 69 in the 1996 set. On closer examination, it will be seen that almost ail of these deletions involved the smallest cities or towns, which in our scheme of territorial re-configuration have been merged with surrounding rural populations. Although these differences are minor, they do make the results for the smallest towns and cities problematic, and care should be taken in interpreting those results. To investigate variations in voting along the urban-rural continuum, we used the same procedure as in the 1993 study, classifying the rayons into city and rural categories, and then subdividing the rural rayons by degree of ruralness. 7 We then re-ordered voting data into each rayon by category of urban and levels of ruralness. We follow the same format for grouping rayons according to urban-rural and by degree of ruralness as we used for the 1993 study. It should be noted here that the data presented in Tables 1 and 3 are in overlapping categories; that is, for the rayons classified as "rural", those in the greater than 75% rural category are also in the greater than 50% rural category, and those in the greater than 50% rural category are also included in the greater than 25% rural category, and so on. This enables one to read across the rural categories, with higher degrees of ruralness to the right. Likewise, city size categories (Tables 2 and 4) overlap; cities of less than 25,000 inhabitants are also included in the less than 50,000 category, and those with less than in the less than 250,000 category, but the category of more than 250,000 is discrete. Party Preference: 1995 Because the 1995 Duma election involved many more parties than the 1993 contest (43 vs. 13). we have also altered the composition and nomenclature of the party groupings that we use to simplify the description of party preference (again, only on the national party list vote). For the 1993 Duma election, we categorized the party groupings as reform. Agrarian-Communist, and LDPR. For the 1995 contest, we changed the designation of the Agrarian-Communist grouping to "left", reflecting the fact that there were two new leftist parties in addition to the APR and KPRF. Also. whereas the LDPR was the only rightist, or nationalist party in 1993, the parties in this segment of the political spectrum proliferated in 1995, so we changed the name of this grouping to "right". The reform grouping of parties retains that name, but the number of parties within the grouping expanded from four in 1993 to nine in Disaggregating the 1995 voting data in our nine-oblast set, we found the same pattern of urban-rural and degree of ruralness differences in party preference as was manifested in the Cities are either discrete administrative entities or aggregations of urban rayons. 8 For a detailed description of the party groupings and our rationale for making them so, see: Clem and Craumer, 1995a and 1995c.

13 Duma election (for an excellent overview of the 1995 Duma election results, see: TsIK, 1995). Specifically, as in 1993, the vote for the reform bloc of parties was consistently higher in urban rayons, and fell in rural rayons along the degree of ruralness slope (Table 1). Also replicating the 1993 pattern, left parties did considerably better in rural areas than in cities (Table 1). Finally, in 1995 the parties of the right showed much less variation across the urban-rural divide and by degree of ruralness, just as was the case in the 1993 contest. There were, however, some important changes in the magnitude of the vote for the different party groupings. The percentage of the vote for the reform parties dropped significantly and that for the left increased dramatically among the units in our data set. This mirrors the nation-wide result, wherein the KPRF and other left parties increased their share of the proportional representation vote by about 12 percentage points, the reform parties dropped by almost four percentage points, and the right fell by approximately three percentage points. Even more significantly, the largest declines in percentage of the vote between 1993 and 1995 for the reform parties was in the urban areas; in 1993 these parties accrued 38.6 percent of the vote in cities within our nine-oblast sample, but by 1995 that share had fallen to 26.4 percent, suggesting a dangerous erosion of the traditional base of support for the reformers. On the other hand, the inroads made by parties of the left in regions of Russia which had favored the reform parties in prior elections and referenda is also reflected in our rayon data; the share of the vote going to the left in the cities of our sample jumped from just 11.3 percent in 1993 to 30.7 percent in Voting for left parties in rural rayons also increased appreciably between 1993 and contributing still further to their overall strong showing. 9 Results of the national party list vote disaggregated by size of city for the December 1995 Duma election generally conform to the pattern of party preference established in earlier electoral events (Table 2). As in our study of the 1993 Duma election, preference for the reform parties is greatest in the largest cities (the ten cities of 250,000 people or more), falls with decreasing size of urban centers, then unexpectedly (i.e.. not previously seen) rises in the smallest towns. 10 On the other hand, as was true in prior elections and referenda, support for the left is relatively weaker in the largest cities, then rises as city/town size decreases. Here again, however, the result in the smallest towns is anomalous, as strength among the left parties dropped in places with fewer than 25,000 inhabitants. Parties of the right enjoy comparatively constant support across city size, excepting again in the smallest towns, where their percentage increased. 9 Our 1995 party blocs were slightly more inclusive than those which we employed in 1993, which contributes marginally to the higher percentages for the left in particular; in other words, some voters who may have been in the "other parties" category in 1993 would be included in the left bloc in The higher left vote is, however, almost entirely a result of much greater affinity for those parties. 10 Again, as there are only four towns with populations less than 25,000 in our sample, the significance of this result is problematic.

14 As was true with the rayons, voter preference for the party blocs changed markedly between 1993 and 1995 by size of city. Thus, in the largest cities (those with 250,000 inhabitants and over) the percentage of the vote for the reform parties dropped by over ten percentage points (43.0 to 32.6 percent) while that for parties of the left soared from about 11 percent to over 28 percent. In medium-sized cities (50,000 to 250,000). the shift from reform to left was even more stunning; in those cities, the left actually outpolled the reformers in 1995 (Table 2). Presidential Elections: witnessed the dramatic conclusion of the multi-year electoral drama in post-soviet Russia with the race for the Russian presidency. In a campaign fraught with implications for the country's future, incumbent president Yel'tsin met nine challengers in a primary round on June 16 and then faced off against Zyuganov in a runoff on 3 July (for an excellent overview of the campaign, see: McFaul, 1996). Nationally, Yel'tsin won about 36 percent of the vote on June 16, followed closely by Zyuganov with 32.5 percent. Of the remaining candidates, Lebed polled a respectable 14.7 percent, with Yavlinskiy (7.5%) and Zhirinovskiy (5.8%) rounding out the top five finishers." Among the nine units of our sample, the distribution of votes by candidate very closely approximated the national result: 36.5 percent for Yel'tsin percent for Zyuganov, 14.7 percent for Lebed, and 6.6 and 7.1 percent for Yavlinskiy and Zhirinovskiy, respectively (TsIK. 1996: ). Support for Yel'tsin was strongest in Sverdlovsk and Murmansk oblasts, while Zyuganov gained a majority of the votes in Bryansk and Kursk oblasts and surpassed Yel'tsin in Saratov and Kemerovo oblasts. Lebed did especially well in Murmansk (with its large military population), while Yavlinskiy reached double-digits in Kirov and Novgorod oblasts (Clem and Craumer. 1996: ). As was noted above, our sample like Russia-contains units of varying population size, which translates into widely differing regional electorates. As we have seen previously, when combined with turnout rates and particular political affinities, these differences can create pronounced spatial imbalances in voting patterns. Thus, within our sample of nine oblasts, 41 percent of all votes for Yel'tsin were registered in just one unit: Sverdlovsk Oblast. Likewise, 73 percent of Zyuganov's votes within the sample set came from his four top units: Saratov, Kemerovo, Bryansk, and Kursk oblasts, in that order. Disaggregating the June 16 figures along urban-rural lines produces exactly the result one would have expected given the experience of prior elections (Table 3). Yel'tsin was the top votegetter in the cities, but his strength ebbed in rural areas. Contrariwise, Zyuganov lagged Yel'tsin in the cities, but outperformed him in the more rural rayons. Yavlinskiy was relatively more popular in cities, just as his eponymous party Yabloko has been, while, as before, Zhirinovskiy's support was 11 As a percentage of valid ballots (TsIK, 1996: 151). 8

15 almost evenly distributed (albeit at a low level) right across the urban-rural continuum. Voting outcomes by city size (Table 4) largely confirm hypothesized trends. Yel'tsin dominated the largest cities, while Zyuganov managed only a quarter of the vote. Yavlinskiy also did best in the largest cities, but Lebed and Zhirinovskiy were most popular in medium-sized cities. The most interesting result among the top five finishers is Lebed's appeal to city voters; it has been suggested that Lebed drew support mainly from an amorphous body of uncommitted voters who were reluctant to support "pure" reform parties and candidates, but also were alienated from the nationalist and left parties (Myagkov, Ordeshook, and Sobyanin, 1997: 162). Given that rural dwellers have traditionally voted left or right, that constituency would have held less promise for Lebed, but there was obviously a previously untapped well of voters in cities for which he was an attractive candidate. Among our sample oblasts, results of the Yel'tsin-Zyuganov runoff matched the national totals almost exactly: 54.4 percent for Yel'tsin and 40.4 percent for Zyuganov. 12 Disaggregated by rayon, the vote tally also conformed to the expected pattern: that is. Yel'tsin did relatively better in the more urbanized rayons and Zyuganov in the more rural areas (Table 3). In fact. Zyuganov actually bested Yel'tsin in all but the most urbanized and, anomalously, the most rural rayons; in the case of the most rural rayons (those over 75 percent rural), the weight of the populous Sverdlovsk Oblast biases the overall result in Yel'tsin's favor. Considering the runoff results by size of city, the dominance of Yel'tsin in urban areas is clear, especially in the largest cities where his margin of victory was almost 25 percentage points (Table 4). Our results also provide some clues as to the question of vote shifting from the June 16 primary to the July 3 runoff, a subject of considerable interest to students of Russian politics (McFaul, 1996: 342; Myagkov, Ordeshook. and Sobyanin. 1997). We correlated the percentage point change in the vote for Yel'tsin by unit between the two rounds with the percentage of the vote for Lebed in round one, and found a coefficient of.769, indicating that units in which Lebed did well on June 16 evinced a major increase in the vote for Yel'tsin on July 3, suggesting a shift in votes from the general to the president. The coefficient for change in the vote for Zyuganov and voting for Lebed is considerably smaller (.361), indicating that the former picked up relatively fewer votes from the latter. This same pattern of movement from Lebed to Yel'tsin was observed by Myagkov, Ordeshook, and Sobyanin (1997: 163) in their exhaustive study of rayon-level voting trends. Our data also show some shifting of votes from Yavlinskiy to Yel'tsin (a correlation of.651) As a percentage of valid ballots. The national figures were 54.4 percent for Yel'tsin and 40.7 percent for Zyuganov (TsIK, 1996: 151). 13 All coefficients are significant at the.01 level.

16 Voter Turnout Voter turnout in the December 1995 Duma election in the oblasts in our sample varied significantly across the urban-rural dichotomy and thence along the ruralness continuum (Table 5). Likewise, turnout differed by size of urban center, but to a lesser extent (Table 6). Both of these patterns were evident in previous elections. It should be noted here that nation-wide there was a ten percentage point increase in voter turnout between the elections of December 1993 and December 1995, reversing a steady decline going back to the 1990 election for the RSFSR Congress of People's Deputies. 14 This increase is reflected in the figures for our rayons: in December turnout among our 303 rayons was 61.6 percent, while in December 1995 it increased to 68.0 percent for the 292 rayons (both figures are higher than the national average of 54.8 and 64.8 percent respectively). In 1995, there was a difference of 11 percentage points in voter turnout between the cities and rural rayons in our data set, and a gap of 15 percentage points between the cities and the most rural rayons (Table 5). The urban-rural split narrowed somewhat from the previous election (it was 12 percentage points in 1993), as did the spread between the cities and the most rural units (18.8 percentage points in 1993). Similarly, the difference between turnout at the extremes of the urban-rural spectrum, that in the largest cities (greater than 250,000) and that in the most rural rayons (more than 75% rural), decreased from 20.3 percentage points in 1993 to 15.3 percentage points in 1995 (Tables 5 and 6). The narrowing of the urban-rural turnout gap was due in all cases to an increase in turnout in urban areas. Despite these increases in urban centers, voters in rural areas continued to turn out at a much higher rate. There was another increase in voter turnout nationally between the December 1995 and the presidential elections of June-July 1996; Russia-wide, turnout rose to 69.7 percent for the first round of presidential voting and then dropped slightly to 68.8 percent for the Yel'tsin-Zyuganov runoff. Turnout among units in our sample also increased to 71.5 and 70.3 percent, respectively. Again. turnout among our sample units was slightly higher than the national figure. For the 1996 presidential races, turnout likewise varied in the expected direction by degree of ruralness and by city size. That is, turnout remains higher in the most rural rayons and relatively lower in the cities. Importantly, however, it should be noted that the urban-rural gap decreased again (down to 7.9 and 8.4 percentage points for the June and July contests, respectively), and the interval between the largest cities (over 250,000) and the most rural rayons dropped to 11.3 percentage points. As was true for the period, all of the change in the urban-rural spread resulted from increases in turnout in urban areas, especially in the largest cities (Table 6). Nevertheless, turnout still correlates negatively with degree of urbanization (-.669 for round one and for round two, both 'Turnout was 78.8% in March 1990, then dropped to 74.7% for the Russian presidential election of June 1991, to 64.5% for the April 1993 referendum, and to 54.8% for the December 1993 Duma election and constitutional plebiscite. 10

17 statistically significant at the.01 level). As we and others have noted previously, the upturn in electoral participation in urban areas, where Yel'tsin has traditionally been strongest, was a major factor in his winning presidential campaign (Clem and Craumer. 1996; McFaul, 1996: 342; Myagkov, Ordeshook, and Sobyanin, 1997: 163). The fact that the percentage of the vote for Yel'tsin in both rounds correlated negatively with turnout in 1996 (-.571 and -.442) and that for Zyuganov positively (.613 and.524) must be taken in historical context; the relationship between reform parties and turnout has been much more negative and that for left parties much more positive in past elections. Differences in Voting Behavior Among Rural Areas A very intriguing aspect of urban-rural electoral geography in Russia is the inter-regional variations which exist among rural areas themselves. 15 That is, after controlling for the degree of ruralness, there remains a very appreciable difference in voting behavior across the country. We illustrate this phenomenon here by adducing figures from the runoff election between Yel'tsin and Zyuganov in July, 1996 arranged by oblast cross-tabulated by rayons according to degree of ruralness (Table 7). As is evident from these data, although the level of support for Yel'tsin generally falls as the degree of ruralness increases, an even more interesting feature of these data is the fact that widely differing behavior occurs from region to region in the same category of ruralness. Thus, in the most rural rayons within our nine-oblast set, the percentage of those supporting Yel'tsin ranged from a low of 28.7 percent in Kursk Oblast to a high of 60.7 percent in Sverdlovsk Oblast. In fact, the level of support for Yel'tsin was higher in the rural areas of Sverdlovsk, Murmansk, Sakhalin, and Novgorod oblasts, than in the cities of Bryansk. Kursk. Kemerovo, and Saratov oblasts. Regional Contextual Effects: Socioeconomic Correlates Clearly, then, there is a place-specific context which conditions political behavior in rural areas. Considering the cases of Sverdlovsk and Kursk oblasts as examples may provide some insights into the influence of the larger regional setting. Sverdlovsk Oblast, part of the southern Urals mining and heavy manufacturing zone, is one of the most highly urbanized (88%, national average: 73%) and industrialized areas in Russia. The oblast center, Yekaterinburg (formerly Sverdlovsk), is one of the largest cities in Russia (1996 population: million), and four other cities in the oblast are over 100,000: Nizhniy Tagil (407,000), Kamensk-Ural'skiy (195,000), Pervoural'sk (137,000), and Serov (100,000). 15 Given this urban-industrial focus, only about nine percent of the oblast's 13 We alluded to this in our earlier article on rayons, but could not explore the subject in depth due to space limitations (Clem and Craumer, 1995b: 469). 16 Goskomstat Rossii, Rossiyskiy Statisticheskiy Yezhegodnik, Moscow: Goskomstat Rossii, 1996, pp

18 population resides in rayons which are more than 50 percent rural; only about 18 percent live in rayons which are 25 percent rural or more. Further, just 7 percent of the population of Sverdlovsk Oblast is employed in agriculture (i.e., only about half of the national average of 13 percent). Thus, there are few areas of a "true" rural nature in the region. On the other hand, Kursk Oblast, in the heart of Russia's Central Chernozem agricultural belt, is about 60 percent urban (i.e., well below the national average of 73%); the city of Kursk itself (1996 population: 442,000), is the only urban center in the region with more than 100,000 people. In contrast with Sverdlovsk Oblast, about 22 percent of the population of Kursk Oblast lives in rayons which are more than 75 percent rural, and about 57 percent live in rayons which are 25 percent or more rural. Approximately 23 percent of the oblast's population work in agriculture (almost double the national average). In other words, the non-urban setting in Kursk Oblast is much more of a "true" rural environment than.is the case with Sverdlovsk Oblast. In their study of the spatial aspect of the Russian Duma election of December. 1993, Slider. Gimpel'son and Chugrov (1994) used factor analysis to develop typologies of regions along two axes: the strength of regions versus the center and degree of economic liberalism (from state control over the economy to decentralized marketization). Of the nine oblasts in our sample, Sverdlovsk, Murmansk, Kemerovo and Sakhalin incline in the direction of economic liberalization, while Bryansk, Kirov, Kursk, Novgorod, and Saratov lean more toward state economic control (Fig. 2, p. 726). Their verbal characterization of these two types suggests that the former group are regions "...that would benefit most from the development of a market economy." and, in the case of Sverdlovsk, also favoring relatively more power for the region vis-a-vis the center (p. 728). The second group falls into the category of regions "...where the local population supports subsidies and continued state regulation of the economy...also included in this quadrant are regions with large rural populations, attracted by appeals from the Agrarians and Communists to retain the system of collective and state farms" (pp ). They also found that "...the distribution of regions along the reform/anti-reform dimension correlates well with several social and economic indicators" (p. 729). Among the indicators cited were average wage and degree of urbanization. 17 Further to this point, O'Loughlin, Shin, and Talbot, in their study of the spatial dimension of the vote for the left and right parties in the 1993 and 1995 Duma elections, concluded that "...region is more than an artifact, but stands for a well-developed geographic consciousness, dating from pre-soviet times, and promoted since 1989 by ethnic-national alignments and contrasting local fortunes in the economic transition" (1996: 382). 17 No coefficient values were provided for the relationship between wages and political tendencies, but the correlation coefficient between the level of urbanization and "statism/reformism" was.61 (Slider, Gimple'son, and Chugrov, 1994: ). 12

19 In previous studies of Russian elections, the April 1993 referendum, and the 1995 constitutional plebiscite, we examined at length the statistical relationship between voting behavior and social structural measures and variables representing economic conditions among the 87 regions of Russia which participated in all of these electoral events (Clem and Craumer, 1993; 1995a; 1995c; 1996). We found strong positive correlations between support for Yel'tsin, his constitution, and the reform parties on the one hand and higher levels of urbanization, education, and white collar employment on the other. Not surprisingly, there is a high degree of co-relationship among these structural variables. At the same time, regions with large rural, agricultural, and older populations tended to vote against the reform parties and measures, and supported the parties of the left. Attempting to ascertain relationships between economic conditions brought about by the post reforms, we also correlated several measures of income; degree of privatization of housing, farming, and industry; unemployment; and cost-of-living against party preference and support for Yel'tsin (Clem and Craumer: 1993; 1995a; 1995c; 1996). Our working hypothesis was: given the uneven spatial impact of reform, those regions which had been relatively less disadvantaged (all regions having suffered economically to some degree) would evince greater support for Yel'tsin and the pro-reform parties, whereas relatively more disadvantaged regions would be less supportive of the president and the reformers. The one variable which consistently tended to support this hypothesis in all of the electoral events was absolute level of income; the other variables, including changes in income, produced weak or inconclusive relationships with voting preferences. As regards the influence of economic conditions on the urban-rural voting gap, we note that income correlated negatively (-.585) with the percentage of the work force in agriculture (1993); the percentage of the work force in agriculture likewise related negatively (-.754) with support for Yel'tsin on the April, 1993 referendum. Further, voter turnout correlated positively with the agricultural work force variable (.676) and with support for the APR (.658) in December These results suggest that rural areas, where the percentage of the work force in agriculture is higher, are typically characterized by lower incomes and, at least partly as a consequence of that, anti-reform voting tendencies. Within the agricultural sector, however, we also found a significant positive correlation (.515) between agricultural wages and support for the Yel'tsin constitution and a significant negative correlation between agricultural wages and support for the APR/KPRF coalition in This finding indicates that rural areas that are relatively better off tend to be more supportive of reform and less inclined to vote for left parties. 18 As we made clear above, one shortcoming of rayon-level analysis is the relative paucity of social and economic data made available at that scale. In our rayon-level study of the 1993 referendum, Duma election, and constitutional plebiscite, we were able to adduce only variables for " All coefficients given above are significant at the.01 level. 13

20 percent rural, age, and education. Yet, the results of this analysis not only confirmed the broader oblast-level outcomes (wherein we employed many more variables), but showed that the statistical relationships intensified because the spatial units were comparatively more homogeneous (Clem and Craumer, 1995b: ). For the 1996 presidential elections we were again able to examine only the same three variables (percent urban, age, and educational attainment) vis-a-vis electoral outcomes but. again, the results proved very interesting. 19 Among the cities and rayons of our nine-oblast sample, the percentage of the population over age 60 correlated positively with the percentage of the vote for Zyuganov on rounds one and two (.491 and.433) and negatively with the vote for Yel'tsin (-.235 and -.396); there was an even stronger negative relationship between older ages and the vote for Lebed (-.557), suggesting that he appealed more to younger and middle-aged voters than to older and retired people. Education also related to candidate preference in the expected direction, with areas of higher educational attainment levels aligning negatively with Zyuganov (-.444 and for rounds one and two) and positively with Yel'tsin (.234 and.399); again, however. Lebed's strength was even more highly associated with the education variable (.556) and the vote for Yavlinskiy also to a large degree (.456). Summary and Conclusion While there is no doubt that urban-rural residence is a major consideration in the analysis of voting behavior in Russia, inter-regional differences in electoral choices suggest that a host of other influences shape the manner in which people vote. Most importantly, the regional context, especially the economic situation as translated through income, appears to play a powerful role, as variations in wealth across rural areas engender different responses to issues, candidates and parties. Other, local concerns, different degrees of sophistication of party organization, the talents and reputations of individual candidates, and other difficult-to-quantify factors likewise enter into the calculus of voting behavior. Finally, several structural variables other than urban/rural per se, including those such as education and occupation which are inextricably bound up with urbanization, must be taken into account when assessing electoral results. 19 For technical reasons, the number of units for this analysis was compressed to

21 REFERENCES Agnew, John A., Place and Politics: The Geographical Mediation of State and Society. Boston: Allen and Unwin, Berezkin, A.V., V.A. Kolosov, M.E. Pavlovskaya, N.V. Petrov, and L.V. Smirnyagin, The Geography of the 1989 Elections of People's Deputies of the USSR," Soviet Geography, 30, 8: October Brustein, William, "The Political Geography of Fascist Party Membership in Italy and Germany ( )," in Social Institutions: Their Emergence, Maintenance and Effects, ed. by Michael Hechter, Karl-Dieter Opp and Reinhard Wippler. New York: Aldine de Gruyter, 1990, pp Clem, Ralph S. and Peter R. Craumer, "The Geography of the April 25 (1993) Russian Referendum," Post-Soviet Geography, 34, 8: , October Clem, Ralph S. and Peter R. Craumer, "The Politics of Russia's Regions: A Geographical Analysis of the Russian Election and Constitutional Plebiscite of December 1993," Post-Soviet Geography, 36, 2: 67-86, February 1995a. Clem, Ralph S. and Peter R. Craumer, "A Rayon Level Analysis of the Russian Election and Constitutional Plebiscite of December 1993," Post-Soviet Geography, 36, 8: , October 1995b. Clem, Ralph S. and Peter R. Craumer, "The Geography of the Russian 1995 Parliamentary Election: Continuity, Change, and Correlates," Post-Soviet Geography, 36, 10: December 1995c. Clem, Ralph S. and Peter R. Craumer, "Roadmap to Victory: Boris Yel'tsin and the Russian Presidential Elections of 1996," Post-Soviet Geography and Economics, 37, 6: June Clem, Ralph S. and Peter R. Craumer, "Regional Patterns of Voter Turnout in Russian Elections, ," in Elections and Voters in Postcommunist Russia, ed. by Matthew Wyman, Stephen White, and Sarah Oates. Cheltenham, England: Edward Elgar, forthcoming. Goskomstat Rossii, (Gosudarstvennyy Komitet Rossiyskoy Federatsii po Statistike-State Statistical Committee of the Russian Federation), Chislennost' Naseleniya Rossiyskoy Federatsii na 1 Yanvarya 1996g (Number of the Population of the Russian Federation on 1 January 1996), Moscow: Goskomstat, 1996a. Goskomstat Rossii, (Gosudarstvennyy Komitet Rossiyskoy Federatsii po Statistike-State Statistical Committee of the Russian Federation), Rossiyskiy Statisticheskiy Yezhegodnik, 1996 (Russian Statistical Yearbook, 1996). Moscow: Goskomstat, 1996b. Hough, Jerry F., "The Russian Election of 1993: Public Attitudes Toward Economic Reform and Democratization," Post-Soviet Affairs, 10, 1: 1-37, January-March Hough, Jerry F., Evelyn Davidheiser, and Susan Goodrich Lehmann, The 1996 Russian Presidential Election. Washington: Brookings Institution Press, Kolosov, V.A., N.V. Petrov, and L.V. Smirnyagin, Vesna 89: Geografiya i anatomiya parlamentskikh vyborov (Spring 89: Geography and Anatomy of the Parliamentary Elections). Moscow: Progress Publishers, Lewis, Robert A. and Richard H. Rowland, Population Redistribution in the USSR: Its Impact on Society, New York: Praeger, McFaul, Michael, "Russia's 1996 Presidential Elections," Post-Soviet Affairs, 12, 4: , October- December Myagkov, Michael, Peter Ordeshook, and Alexander Sobyanin, The Russian Electorate, ," Post-Soviet Affairs, 13, 2: , April-June O'Loughlin, John, Colin Flint, and Luc Anselin, "The Geography of the Nazi Vote: Context, Confession and Class in the Reichstag Election of 1930," Annals of the Association of American Geographers, 84, 3: , September O'Loughlin, John, Michael Shin, and Paul Talbot, "Political Geographies and Cleavages in the Russian Parliamentary Elections," Post-Soviet Geography and Economics, 37, 6: , June Orttung, Robert W. and Scott Parrish, "Duma Votes Reflect North-South Divide," Transition, 2, 4: 12-14, 23 February Petrov, Nikolai, "Analiz rezul'tatov vyborov 1995g v Gosudarstvennuyu Dumu Rossii po okrugam i regionam (Analysis of the Results of the 1995 Russian State Duma Elections by District and Region)," in 15

22 Parlamentskie \ybory 1995 goda v Rossii (The Parliamentary Elections of 1995 in Russia), Moscow: Carnegie Endowment for International Peace, 1996, pp Slider, Darrell, Vladimir Gimpel'son and Sergei Chugrov, "Political Tendencies in Russia's Regions: Evidence from the 1993 Parliamentary Elections," Slavic Review, 53, 3: , Fall TsIK (Tsentral'noye Izbiratel'naya Komissiya Rossiyskoy Federatsii-Central Electoral Commission of the Russian Federation), Vybory Deputatov Gosudarstvennoy Dumy 1995 (Elections for Deputies of the State Duma 1995). Moscow: Ves' Mir, TsIK (Tsentral'noye Izbiratel'naya Komissiya Rossiyskoy Federatsii-Central Electoral Commission of the Russian Federation), Vybory Prezidenta Rossiyskoy Federatsii 1996 (Elections for the Presidency of the Russian Federation 1996). Moscow: Ves' Mir,