Giorgian Guțoiu, Urban Administrative Fragmentation and Its Consequences on Electoral Behaviors at Romanian National Legislative Elections. A Case Study on BucharestEuropean Quarterly of Political Attitudes and Mentalities ISSN 2285-4916 ISSN-L 2285-4916 Volume 9 Issue No.4 October 2020 pp.14-37.Published: 25 October 2020
Urban Administrative Fragmentation and Its Consequences on Electoral Behaviors at Romanian National Legislative Elections.
A Case Study on Bucharest
Giorgian GuțoiuFaculty of Social Sciences and Humanities"Lucian Blaga" University of SibiuRomania
Date of submission: September 29th, 2020 Date of acceptance: October 20th, 2020 Date of publication: October 25th, 2020
AbstractMany of the present day major cities have their territory partitioned into administrative subdivisions for a wide range of local governance purposes. Partisan local elections are held for electing the politicians into public administration offices of the subdivisions, which may lead to the creation of local partisan conflicts within each unit. In this paper, I analyze the influence of this local political context on the electoral behaviors formed at the national legislative elections to see if the local political context in the subdivisions determines a lack of democratic representation by rendering the socio-spatial conflicts of the city irrelevant. I explore this on Bucharest, a large city of nearly two million, at the national legislative elections between 2000-2016. The methodology is drawn from the subfield of electoral geography, as I spatial analyse the geographic clustering of electoral behaviors in the six subdivisions of Bucharest. The results in Bucharest show the capacity of the subdivisions to influence voting decisions through the local political context when certain conditions are met.
Keywords: urban administrative subdivisions, electoral geography, electoral behaviors, Bucharest
In the context of an increasingly complex urban life, many of the present day major cities are partitioned into administrative subdivisions for a wide range of local governance purposes, like taxation, sanitary services, transportation, education, sewage or water facilities. Administrative partitioning is present both within cities, but also at the upper scale of metropolitan areas. In this paper, I am interested in the subdivisions within cities, since this administrative fragmentation received much less attention in recent decades than that of metropolitan areas. Regularly, the degree of administrative fragmentation goes along with the size of the cities. Larger cities are more fragmented. However, there are significant differences worldwide. In the United States, it is quite common for small cities to have several administrative subdivisions. In the Global South, or in Central and Eastern Europe, generally, the major cities or capitals are partitioned. In Western Europe, fragmentation is quite common for medium-sized cities as well. In this paper, I am interested in the political fragmentation generated by this administrative partitioning. For this, I explore the influence political fragmentation arised from administrative partitioning has on the electoral behaviors within large cities. I research this topic through a case study of the legislative national elections in Bucharest, Romania’s capital, between 2000-2016. While today's urban spaces undergo increasingly dynamic and rapid socio-cultural micro-transformations, this field of political and administrative fragmentation in cities remains very poorly explored.
Previous studies have shown that there is an influence of local political context on the electoral behaviors formed in that place (Johnston and Pattie, 2006). However, studies in this area have been particularly concerned with elections in localities or constituencies, while the urban level has not yet been fully investigated. One way in which the local political context influences electoral behavior is through the existence of a partisan majority or a strong local political leader. In an older study of the US cities it was observed that there is no precise and clear influence the administrative fragmentation of cities has on political participation (Lowery and Lyons, 1989). In this regard, I will operationalize the local political context in subdivisions by identifying the partisan affiliation of mayors in local institutions. For this matter, I set the field of research to cities fragmented into subdivisions where partisan elections are held for local public offices. The present case study has a decisive exploratory nature. There is, however, a precise objective to the study, namely that of understanding the influence of the local political context from subdivisions on the electoral behaviors in the general national elections.
In the realm of political decision making process, questions of urban governing and institutional hierarchies are currently dominated by the theme of large metropolitan areas because, in the current stage of globalization, the success of urban spaces is decisively determined by their performance in an ultra-competitive economic space, henceforth policy relevant actors seek to create these grand regions of clustered and integrated urban spaces (Brenner and Theodore, 2002; Frisken and Norris, 2001). Also, academics concerned with democracy and urban governance were mostly focused on the new institutions created at the metropolitan level (Kubler, 2012; Buser, 2013; Zimmermann, 2014). Little attention is given to the political consequences generated from the administrative fragmentation of today's increasingly diverse social and cultural cities. Through the present paper I propose a contribution on this topic. I shall analyze the influence of the political life in the sub-divisions on the electoral behaviors at the national legislative elections in Bucharest between 2000-2016. For this, I employ spatial econometrics to quantify the clustering of electoral performances in the geographical space of Bucharest’s subdivisions. I discuss the spatial concentration of electoral behaviors in subdivisions by reference to the local political context to highlight and explain the empirical patterns. Through the approach and topic, this paper also contributes to the growing literature on Romanian urban electoral geographies (Guțoiu, 2018, 2019).
The following section presents a framework for understanding electoral behaviors in their geographic context. Afterwards, the focus is shifted to the case study. Firstly, I discuss the Romanian electoral politics and the ways in which local mayors tend to influence local electoral behaviors. Secondly, I present the administrative partitioning of Bucharest into its six subdivisions and describe their socio-spatialities. Next, I discuss the local political context, referring to the distribution of mayors and their partisan affiliation, in the subdivisions of Bucharest at the local elections from 2000-2016. After, I describe the statistical tools used for electoral spatial data analysis. Following these, I explore the clustering of electoral performances in Bucharest at the level of subdivisions and discuss the geographic structuring in relation to the local political context and socio-spatiality. The final conclusions close the paper.
2.Understanding electoral behaviors in their geographic context
The approach used in this paper is underpinned by theorizations and empirical advances from the subdiscipline of electoral geography, in which the research of voting spatiality is the central topic (Leib and Quinton, 2011). In electoral geography there are two main approaches to the spatiality of voting. One deals with the simple description of the spatial distribution of voting. Another area, which also fundaments the approach in this paper, is concerned with the relation between the voting decision and its geographical context. In the last decades, within the subfield of electoral geography a robust volume of works has been generated to help explain the mechanisms of this influence. In the following paragraphs I focus my attention on these developments.
In this approach, the geographical context is not only an empty space where the electoral behavior is formed independently, but it's intrinsically linked to the construction of the voting decision (Agnew 1990). However, scholars of electoral geography do not agree on the precise set of mechanisms by which the geographical context counts for the electoral behaviors (Van der Wusten and Mamadouh, 2014). One approach in electoral geography explains the contextual influence through microsociologies, where the voting decision is influenced by the social networks that constitute the geographical context (Pattie and Johnston, 2000). Another approach, developed within American electoral geography, adds to this microsociological explanation also a geographic perspective (Agnew 1987 and 1996). These two approaches are complementary (Johnston and Pattie, 2006).
John Agnew (1987 and 1996) developed theoretical arguments for understanding the relationship between voting and its geographical context. Agnew argues that the voting decision is formed within a context where microsociologies and structures of everyday life meet with social, political, economic and cultural multiscalar processes. A major role is played here by political parties. Electoral behaviors are directly linked to the political agenda set by parties. The society and the state vary geographically, also the political parties acting as mediators between the two develop by influencing and being influenced by space. The geographic context matters for voter. In the geographic space voters meet the parties. Also, the political parties are not merely electoral vehicles, but instead they channel expectations from voters, and are therefore evaluated according to their potential to distribute resources, so finally the electoral geography becomes mostly the spatialization of who get’s what? (Shin and Agnew, 2008: p. 52). Because, today, in post-industrial democracies, the old partisan attachments based on social classes vanished, and the political competition has become multidimensional and value-based (Evans, 2017), the analysis of electoral behavior can be significantly improved through a contextual perspective that captures the dynamic contemporary democratic world.
The influence of political life from an administrative subdivision on electoral behaviors in that subdivision I conceptualize it as a contextual influence of the geographic context on the voting decision. Here, political parties play a primary role. Studies of electoral geography based on a social construction of the space described the relationship between electoral mobilization and political parties as one mediated and influenced by the geographical context. For this reason, before the spatial analysis of voting, I discuss the operationalization of geographical context, by referring to the political and socio-spatial processes relevant for understanding electoral behaviors in the geographical context of Bucharest’s administrative subdivisions.
3.Mayors and local electoral politics in Romania
Understanding the influence that a political context has on electoral behaviors within a space starts with the description of the actual political context. For our case study on Bucharest, the political context of interest is located within the six subdivisions, where elections are held every four years for the subdivision’s mayoral and local legislative seats. In this section I present the mechanisms through which Romanian mayors influence voting decisions at national elections. In this regard, I describe the literature on electoral mobilization generated by local Romanian politicians and mayors.
This influence is determined by the nature of Romanian democracy, marked by low ideological institutionalization of party system (Gherghina and Jiglau, 2011), the high level of fluidity of the party system (Soare et al., 2013), the weak partisan identification (Gherghina, 2015) and a strong personalized competition (Gheorghiță, 2014). In Romania, mayors are elected by popular vote. Mayors are the executive authority, while a local council elected through proportional representation is the legislative body. Usually, Romanian citizens have more trust in the mayor than in the MPs (Tufiș, 2008), and thus are more likely to be influenced by their actions for mobilization. Ties with local politicians are stronger, as the MPs are perceived rather more corrupt, distant and distrustful. Local electoral campaigns are carried out with the high involvement of mayors, who are usually also the leaders of local party branches. These local influential politicians are rewarded with important revenues for their major contribution to their parties electoral success (Roper, 2006; Gherghina and Chiru, 2013). Such revenues can become a factor in building local loyalties. Private contributions are another important part of electoral mobilization resources (Gherghina and Volintiru, 2017). Also, Romanian politicians who are serving as mayors, use the office to increase their political capital to facilitate the reelection (Chiru and Gherghina, 2012). In Romania, the number of reelections for a mayor are unlimited. In a country with weak democratic institutions and doctrinary competition, such as Romania, poorer voters benevolently keep corrupt politicians in power because of the perception that they can derive material benefits from corrupt leaders exploiting private or public resources (Manzetti and Wilson, 2007). In addition, corruption also exists as vote buying (Gherghina, 2013).
4.Administrative subdivisions and socio-spatialities in Bucharest
Bucharest’s subdivisions are named sectors (sectoare) and are officially named with numbers ranging from 1 to 6 and resemble the geometric shape of a circular sector (see Figure 1). The six sectors are pretty much similar in their extent and shape. In fact, the sectors are also similar in terms of socio-spatiality. In Table 1 we see that the surface and total number of inhabitants vary little between the six sectors. With the exception of sector 1, Bucharest is divided into subdivisions of almost equal sizes. Sector 1 has a larger area due to the extensive green spaces. The socio-spatialities differ little between the subdivisions. This is due to the circular sector shape of the subdivisions and the radial-concentric historical and geographic development of Bucharest.
Bucharest’s subdivisions and types of housings.
Population and area for Bucharest’s sectors.
Inside them, each sector has strong differences. The sectors were created in 1979 under the direct autocratic leadership of Nicolae Ceaușescu, with the purpose of remodelling Bucharest through uniformizing its socio-spatial by administrative fragmentation. During the totalitarian regime, Bucharest was reshaped by the political and economic exploitation of the state (Cavalcanti 1997). All urban transformation was part of state policy to standardize the urban and social space (Zarecor, 2018). The socio-spatiality of post-1989 Bucharest is mostly the consequence of state socialism transformations. About 80% of the houses in Bucharest are socialist era apartments. Through the ubiquity of its durable physical concrete constructions, the socialist city preserved its urban legacy. However, the socialist cities were not some monoliths of concrete blocks of flats. Despite the regime's official intentions to achieve uniformization, socialist cities were displaying solid patterns of socio-spatial segregation (Marcińczak et al., 2014). These patterns were either inherited from the bourgeois era, or they were the result of preferential housing allocation during socialism. Also, the quality of socialist housing and access to public services were not equally spatially distributed. After 1990, Bucharest develops under the neoliberal project of urban development, so that the socio-spatial differences inherited from the state socialism era are strengthened.
The large socialist housing estates are located mainly outside the central historical city, consisting of pre-WW II single-family dwellings, villas, and small blocks (Figure 1). High-quality socialist blocks are also clustered in central city on main boulevards and in central-southern area transformed by Ceausescu. The central city has also small pockets of spatial violence between the wealthy and the vulnerable groups (Armaș and Gavriș, 2016), however, as a general feature, the historical city concentrates the high-value land (Rufat and Suditu, 2008). As each sector comprises a segment of the historical central city, the situation creates strong socio-spatial contrasts within the subdivisions.
The largest areas of socialist collective housings are located in sectors 3,4 and 6. Smaller socialist housings are located in sectors 1, 2 and 5. Within all six subdivisions, these socialist developments oppose to the older pre-WW II city. However, the socialist city is not monolithic. Significant spatial differences exists in terms of quality and equity of access to public services. Social filtering based on housing prices developed in the present neoliberal era also as a result of these conditions. The poorest and least desired areas on the housing market in Bucharest are the self-built neighborhoods and the low quality socialist blocks. High quality blocks are those in central area, while medium quality blocks are clustered on main roads or within the microrayons from sectors 3 and 6 (Maxim, 2018).
Alongside the socialist era fabric, within the area outside the historical space there are also former neighborhoods of pre-WWII villas, located mainly in the sector 1 which is wealthier and with a higher share of low-density houses and green spaces than other subdivisions. Vulnerable populations are predominantly found at the spatial peripheries of the city (excepting sector 1). The largest area comprising poorer groups is located in sector 5, in the Rahova and Ferentari neighborhoods, dominated by low quality socialist era blocks or self-built homes. In sector 5, a strong contrast exists between the poorer areas and the pre-WW II villas of Cotroceni (the most expensive neighborhood). The vast majority of those near to the median income are located in the large socialist estates of sectors 3, 4 and 6. On the outer ring, on small pockets, former semi-urban and quasi-rural peripheries during socialist era are now transformed into low-density neighborhoods for the suburbanites.
Socio-spatial contrasts are present both within and between the sectors. Within sectors, the strongest contrasts exist between the historical central area and the space outside it. The pre-WW II, socialist, and neoliberal urban transformations and the socio-spatial contrasts created from these developments all lack political representation at the level of Bucharest’s subdivisions.
5.Local political contexts in Bucharest’s subdivisions
In this section, I describe the political context of Bucharest’s subdivisions through a longitudinal perspective during the years 2000-2016. The purpose is to understand the political contexts and conditions developed within subdivisions that may influence electoral behaviors in national legislative elections. The local political contexts are those created following the local elections, which are held during the first half of the year, a few months before the legislative elections. In this regard, I use for analysis the results at the elections for mayoral seat. Also, supplementary, I discuss the distribution of seats in the subdivisions local council. Since the Romanian legislation does not limit the number of terms a mayor can have, politicians can build and consolidate strong loyalties as mayors throughout many years. At the 1996 and 2000 elections no mayor was reelected. This prevented the appearance of local loyalties within Bucharest’s subdivisions. The mayors that won in 1996 and 2000 were even from other parties than their predecessors. Only starting with the elections of 2000, a political capital is beginning to form at the top of the administrations in subdivisions. We can see this phenomenon in Table 2 which shows the partisan affiliation of politicians who have won mayor or local councilor seats at the local elections in Bucharest between 2000-2016. Annex 1 comprises the name of parties and alliances in Romanian and English, since for text flow reasons in the body of the paper I use only the parties and alliances acronyms. Also, in this paper, I make no reference to the names of the local mayors as I believe this is not relevant to the arguments made in this paper as the constant name references would only hinder the text flow.
At the 2000 elections, the mayoral seats and pluralities in all seven councils (Bucharest and its subdivisions) were won by the PDSR. The landslide victory was achieved by the PDSR because of its position as party in opposition which gained much support as a consequence of the protest vote against the CDR, the incumbent alliance during 1996-2000 and the main actor in Bucharest’s local politics during the 90s. The PDSR won proportional shares of seats in all subdivisions. As there were no solid political loyalties within sectors, the parties performances didn’t vary much between sectors. The mayoral elections for Bucharest were won by the charismatic Traian Băsescu (PD).
In 2004, the PNL-PD Alliance led by Băsescu, who is reelected as city’s mayor, won most of the mayoral mandates and pluralities in local councils within subdivisions. The alliance consisted of two partners. The PNL won mayoral seats in sector 1, 4 and 6, and the PD in sector 3. In sectors 2 and 5, loyalties were formed, as the PSD (PDSR) mayors were reelected.
At the 2008 elections, five mayors of sectors were reelected. In sector 3, the PDL won the mayoral seat and a large majority in the local council. In sector 6, the PDL candidate was reelected as mayor, having won in 2004 as PNL’s candidate. In sectors 2 and 5, mayors from the PSD were again reelected, while the party won plurality of seats in the local council. In sector 1, the mayor from the PNL was reelected, while the PDL won plurality in council. In sector 4, a candidate from the PC (a small party linked to the PSD) is elected from the first time. At the city level, the PDL won a plurality in the council, while the mayoral seat was won by Sorin Oprescu, an independent candidate close to the PSD. At the 2008 elections, with the exception of sector 4, in each subdivision there were solid loyalties towards one politician or political party.
The main feature of the 2012 elections was the massive protest vote against the PDL, the incumbent party during the financial crisis. In this context, the USL, alliance formed by PSD and PNL, scored a landslide victory in Bucharest, winning five mayor seats and majorities in every council. Excepting the candidates from the PDL, all other mayors were reelected.
In 2016, the PSD in alliance with a small party (the UNPR) achieves a landslide victory, winning all seven mayoral seats and pluralities in all councils. However, it is worth mentioning that five of the elected mayors won their first term. Toghether with the local sympathies towards their politicians, a major part of the PSD’s succes in 2016 was caused by the lack of any solid opposition from the other parties. The main opposition towards the PSD in Bucharest was the USB, a newly established urban regional party in Bucharest.
In this section, the intention was to describe from a longitudinal perspective the local political contexts in Bucharest’s subdivisions, in order to understand local political loyalties and sympathies towards politicians or parties. There appears to be significant patterns of temporal evolution. Support towards local politicians or parties grow up to the 2008 local elections when they are the strongest, as five mayors from three different parties won the reelection. Each sector had a dominant political sympathy. Starting with the 2012 elections, the degree of reelection gradually decreases. Because the local political capital diminshes and some strongholds dissapear, at the elections of 2012 and 2016, the city is dominated by actors that extended their power relatively proportionally throughout the subdivisions.
Local elections' results in Bucharest.
6.Description of data and statistical tools
The empirical part of the paper implies the spatial analysis of electoral data in order to identify geographic clusters of support. In this section, I discuss the methodology used for spatial analysis. The electoral data consists of votes (percentages) scored at the legislative elections between 2000-2016 by the political parties of interest. Also, the selection of these parties I describe further in the current section.
Clusterization of values for a variable within geographic space is commonly known as spatial autocorrelation or spatial dependence (O’Sullivan and Unwin, 2003) and is measured through spatial econometrics, which acknowledges and integrates within calculations the correlation between observations determined by their proximity within space. In recent decades, spatial econometrics tools have gained recognition as necessary methods for understanding political behaviors within their geographic context (O’Loughlin, Flint and Anselin, 1994; Shin and Agnew, 2011). In spatial econometrics, the data are analyzed by reference to their geographic position. The electoral data I employ are georeferenced through the geographic coordinates of polling locations.
One classic instrument for measuring clustering in geographical space is Moran's I (Getis, 2010), which is defined as:
w is an unit in a row-standardized spatial weights matrix (W) that encapsulates the spatial proximity between observations (i.e. 1=connection; 0=disconnection), and
x is the vote share for party of interest.
The matrix I employ for this study summarizes three connections for each observation, as a result of distance-based definition of the three most proximate neighbors. Actually, Moran’s I is an index that measures the relations between observations in space and their spatially lagged correspondence, which is the weighted average of values from surrounding observations. Values of I range from -1 (dispersion) to 1 (clusterization). To calculate the statistical significance, a z-score can be computed by producing randomly distributed samples (Anselin, 1988). To calculate Moran’s I and z-score I employ the free software package GeoDa, vers. 1.14.0. The index’s value expresses the degree of global spatial autocorrelation of values within the geographic space of interest. For this paper, the value of Moran’s I expresses the spatial autocorrelation for voting share (percentages) of political parties at the level of Bucharest.
In order to identify local clusterization I use Local Moran’s I statistic, proposed by Luc Anselin (1995) as a Local Indicator of Spatial Association.
The formula for Local Moran’s I is:
z is measured through standard deviations from the average, and the inference is computed through a similar randomization method as for Global Moran’s I, except now it is computed in turn for each observation.
Local Moran’s I is a statistic computed for each observation and measures the degree of clusterization surrounding that observation.
In GeoDa, for each observation, a Local Moran's I is calculated, its statistical significance is verified and one of the following four types of spatial association is identified:
1) clusters of high values surrounded by high values,
2) clusters of low values surrounded by low values,
3) outliers of low values surrounded by high values, and
4) outliers of high values surrounded by low values.
For the analysis in this paper, I employ cluster maps computed in GeoDa, with statistical significant spatial association regimes identified at a level of p=0.05. After describing the statistical tools used in exploring the electoral geography of Bucharest’s subdivisions, the rest of the article continues with the spatial analysis of electoral data.
7.Exploring electoral behaviors in Bucharest’s subdivisions at the legislative elections during 2000-2016
I have argued in the previous sections how the administrative fragmentation of Bucharest enables the developing of certain local political contexts that can influence behaviors at other national general elections. As the city becomes increasingly divided from partisan politics formed within its administrative subdivisions, the political representation at legislative elections will become less focused on socio-spatial conflicts and more on partisan conflicts that transpose conflicts between political or administrative elites as well as loyalties towards these elites. In this section, I move further to the analysis of relationship between local political contexts in sectors and the electoral behaviors expressed at the legislative elections. For this, I shall measure the degree of geographical clustering of electoral behaviors in Bucharest and its subdivisions. This clustering is further discussed in relation with the local political context and the socio-spatial contrasts.
Table 3 gives descriptive statistics for parties and alliances of interest. The results are aggregated at the level of subdivisions, Bucharest and the national. The table depicts actors with at least 10% of votes in Bucharest at the legislative elections. As an exception, the PNL (9% in 2000) is also placed here because of its importance in national and local politics. In the present section I shall spatially analyse the electoral results for these actors. Because the legislative elections are held a few months after the local elections, most of the actors in Table 3 are under the same name in Table 2.
However, the reader is encouraged to use also the Annex 1 to better understand the actors in the context of these high fluidity Romanian politics.
Descriptive statistics for the national legislative elections in Bucharest and the national level.
Results in Table 3 are provided mainly as points of reference and a starting point for the following spatial analysis. In this regard, the standard deviation values do provide an image on the variation of performances between subdivisions and it features a clear longitudinal pattern. At the 2000 elections, the standard deviations were low for the PDSR-PUR-PSDR (alliance formed from the dominant PDSR party and other small two parties), the PD and the PNL, and at an average value for the PRM and the CDR. The PDSR who won the elections with a consistent margin had a largely proportional performance between sectors. In 2004, the competition was fought by two major actors. For the PSD, the dominant party of Romanian politics, who participated again in alliance with the small party of the PUR, the standard deviation increases to 1.23. The value is much higher for the other main actor, the Alliance of the PNL and the PD, namely 3. The PSD+PUR won the elections at the national level, however, in Bucharest, the PNL-PD won by a large margin of 17%. In 2008, all the three actors, the PSD+PC, the PDL and the PNL have electoral geographies with major variations between sectors. Values are at an all-time high for the period of interest for each of the three actors: the PSD+PC 6.23, the PDL 7.60 and the PNL 4.35. The standard deviations decrease in 2012, when the USL alliance achieved a landslide victory throughout the entire city. Its scores were proportionally distributed between subdivisions. Although the 2016 standard deviations are higher than in 2012 they do not reach the 2008 level. The analysis of standard deviations and aggregate scores shows a measure of variation between sectors, however it does not help us identify the geographic clustering. For this reason, the paper continues with the spatial analysis of electoral data.
I begin the spatial analysis with the Moran’s I, which gives a global measurement for geographic clusterization of electoral scores at the level of the entire city. Moran’s I and z-scores for candidates of interest are displayed in Table 4.
Note that a Moran’s I value of 1.0 indicates perfect positive spatial autocorrelation, meaning that each observation can effectively be predicted by the average of nearby observation. For the positive spatial autocorrelation to be positive, the z-score must be higher than 1.96. The values are analyzed in the following paragraphs.
Global Moran’s I and z-score in Bucharest.
The degree of global spatial autocorrelation varies between candidates. The highest degree of geographic clusterization we find at the elections of 2008. The value of Moran’s I signals a high degree of clusterization for each candidate, the PSD+PC, the PDL and the PNL, respectively. The 2008 legislative elections displayed, at both national and local level, a competition that was close to the center and had weak polarization. The elections generated a quite low level of interest from the public, also producing the smallest turnout (39%) in all legislative elections held after 1990. If the electoral mobilization is generated mainly by the national level political agenda, an electoral campaign that does not polarize the public space and does not stresses subjects of social conflicts should create a weak clustered electoral geography. However, the 2008 electoral geographies in Bucharest were heavily clustered. The hypothesis of the paper suggests that this clusterization may be the result of local political contexts in Bucharest’s sectors, as the city’s subdivisions at that time had mayors from three different parties and most of the subdivisions were strongholds for politicians and parties at the top level of local public administration.
At the 2000 elections, high degree of clusterization is present only at the CDR and the PRM. Their electoral support came from well-defined social groups. The CDR, was an alliance formed around its main element, the PNTCD which was at the time the main party in the governing coalition. The alliance was struck by a strong protest voting and relied only on its electoral core, namely those relatively high educated or wealthy. The PRM was a populist party that relied on protest votes mostly from those vulnerable. The PDSR-PUR-PSDR, an alliance with its main element, the dominant PDSR, won the 2000 elections by relying on protest votes and a centered stance. In Bucharest, where the alliance also won with a large margin, did it with no significant loyalties throughout the subdivisions. This may explain the low-to-moderate degree of global clusterization of the PDSR-PUR-PSDR electoral geography in Bucharest. At the 2004 elections, values of Moran’s I indicated that there is an average degree of clustering, as the competition was polarized between the PSD+PUR and the PNL-PD. The first was stronger in the rural areas and the poor urban spaces and relied more on those disadvantaged. The latter relied on those generally more educated and younger. At that time, unlike in 2000, there were also some loyalties formed towards local politicians in the subdivisions, as the mayors in sector 2 and 5 were reelected earlier that year. Also, the city was divided between these two alliances, since the PNL-PD also won some mayoral seats in the local election. As the loyalties towards local politicians and parties gradually increased between 2000-2008 so did the degree of clusterization.
After 2008, the degree of clusterization decreases, but remains rather at medium level. In 2012, both the local and legislative elections were won by the USL. Its success was due to the economic vote against the PDL (ARD) which implemented many austerity policies during its incumbency between 2009-2012 and was left in Bucharest only with its hardcore electorate consisting of relatively educated, wealthy and young, but with very little potential to mobilize with the help of local politicians. As a result of these clearly-defined patterns of support, the ARD’s electoral geography was strongly clusterized. In comparison, the USL’s electoral geography was rather moderated clusterized. At the local elections, the USL achieved a landslide victory across the subdivisions. Its rhetoric attracted diverse social segments, as the alliance was formed by two parties with different electoral bases, the PSD targeting the poorer and the PNL those more educated. At the 2016 legislative elections, high clusterization is found at the two main parties in Bucharest, PSD (I=0.7) and USR (0.6), while the third ranked party, the PNL, had a low-level clusterization (0.23). As the recent previous elections, the 2016 competition attracted again little interest from the voters (39% turnout). However, the presence of the USR, a new party targeting the higher educated and younger strata, slightly radicalized the scene. In 2016, the PSD won all seven mayoral seats and pluralities in all councils. Yet, as we seen from the descriptive statistics, both the scores of the PSD and the USR had major variations between subdivisions. This was a departure from the 2000 and 2012 victories of the PSD, when its scores were relatively proportionally distributed.
Moran's I analysis showed that high spatial autocorrelations exist at the 2008 elections. Medium-to-high levels are present in 2012 and 2016. The 2000 elections marked the lowest spatial autocorrelation. The 2000-2008 saw the gradual increase in spatial autocorrelation, at a time when the mayors won their first (and second) reelection. However, the Global Moran does not show the local patterns of support, hence a final conclusion on the spatial autocorrelation and the local political context should be stated after the local analysis, which follows next.
For text flow reasons, LISA maps are displayed in Annex 2. The maps show clusters that are significant at 0.05 level and can be used as benchmark for comparison. In the following analysis, I shall quantify only the clusters of high-high values or low-low values which have all of its four observations in the same sector. This option better isolates the spatial contours of local political contexts and behaviors subdivisions. The number of groups with central location and neighbors in same sectors are listed in Table 5. The majority of locations are placed in the same sector as their three neighbors.
Overall distribution of groups of polling locations and neighboring polling locations within the same sector
The quantified clusters of interest are shown in Table 6. Within each subdivision and for every candidate the share of clusters (high-high or low-low) is computed from the total groups of central locations and their neighbors located in the same sector. The table shows a degree of local spatial autocorrelation for each party or alliance in relation to the geographic space existing in the respective sector and helps to identify the geographic clustering of performances in Bucharest’s administrative subdivisions. As a rule of thumb, I consider weights of 50% or more to be eloquent for revealing an important concentration.
Share of clusters comprised of central location and neighboring location located within the same sector computed from the total number of potential clusters within the respective sector.
At the legislative elections of 2000 and 2004, there was no major clusterization that can be explained as the result of a local political context. Extended spaces of clusters in 2000, namely those high-high for the CDR in sector 1 and for the PRM in sector 5, those low-low for the CDR in sector 5, low-low for the PRM in sector 1 and in 2004 those low-low for the PNL-PD in sector 5, are best explained by the distribution socio-spatial in these subdivisions and the clearly-defined rhetoric of the political actors. In 2004, the PSD+PC was the only actor with reelected mayors, in sector 2 and 5. However, the alliance does not have major clusterization located in these two subdivisions. The local and legislative elections 2004 were won in Bucharest with a large margin by the PNL-PD alliance. The competition in Bucharest was polarized by the national political arena. The presidential elections of 2004 were held simultaneously with the legislative ones, and the mayor of Bucharest, Traian Băsescu, the candidate from the PNL-PD side, wins the run-off. His candidacy acted as a strong bandwagon effect for its alliance. This was the context in which the influence of the PSD mayors reelected in 2004 was substantially reduced. This may also explain why the sympathy for PNL-PD was distributed relatively proportionally between sectors, although the alliance had won four mayoral seats earlier that year. Unlike 2000 and 2004, the influence of the subdivisions on the electoral behaviors in Bucharest was substantially inforced at the next legislative elections.
As shown in Table 6, electoral support at the 2008 legislative elections was highly clustered within Bucharest’s subdivisions. Each of the three actors of interest scored much better performances in the sectors where they had mayors. For the PSD there are 67% clusters of high-high values in sector 5 where the local mayor was reelected for its third term, and 47% in sector 4 where the mayor was elected as a candidate of the small party PC. The PNL achieved its best performances in sector 1 (83% high-high clusters). The PDL concentrated its best scores in sector 3 (67% high-high clusters). Clusters of high support are missing in sector 2 and sector 6. For the PSD+PC the best scores are clustered in sector 4 and sector 5. It is here that on a wide geographical area both the local political context and the socio-spatial is favourable to the alliance. For this reason, the PSD+PC does not concentrate its high performances in sector 2, although the local mayor was constantly reelected. In sector 6 there was no geographic clustering. The PDL scored in this subdivision performances above its average, yet they were far from those recorded in sector 3. In sector 6, the mayor only recently switched sides to the PDL, so the local electoral base of its new party was not so strong as in sector 3. The strong geographical clustering of electoral behaviors within Bucharest subdivisions at the 2008 elections can be attributed to the local political contexts within the administrative divisions of the city. At that time, the city was partitioned by partisan arrangements within sectors, as almost each sector had strong loyalties towards one politician or political party. Also, each of the three actors, the PDL, the PNL and the PSD+PC, had a stronghold in at least one subdivision. Yet this context was also enhanced by some features of the political competition. The 2008 elections were not polarized at the national level, they lacked ideological battles and generated the lowest turnout in all post-1990 legislative elections.
At the following elections of 2012 and 2016, the clustering within subdivisions is reduced to lower levels. Some clusters of high support are present in sector 2 for the ARD alliance (43%) which included also the UNPR, the small party of the local mayor, and in sector 5 in the USR’s electoral geography (60%). However, the general clustering within the sectors is reduced, in comparison with previous elections. Now, the electoral scores are rather dispersed throughout the city. The factors I used to explain the clustering at the 2008 elections disappeared or were much weaker at these elections. The same parties and alliances that won with a landslide victory the local elections were also the victors with large margins in almost all subdivisions. In 2012 and 2016, Bucharest was not divided into strongholds of different parties as it was in 2008. Moreover, the loyalties created in previous electoral cycles disappeared after 2008. For example, in sector 3, the legacy of the PDL’s 2008 major success does not count almost at all for the ARD in 2012 or the PNL in 2016 (the two parties merged in 2014). In 2016, the clusters are generated by the socio-spatial of the city and the national rhetoric of parties, since the USR mobilizes much stronger in the richer, educated and younger areas, whilst the PSD builds an electoral geography with hotspots of support in the poorer areas. Also, the 2012 and 2016 elections were more polarized at the national level and had stronger ideological dimensions. At these elections, the local political context was less important than in 2008.
In this article I have studied the influence administrative fragmentation of a big city has on the electoral behaviors in that city during national legislative elections. The influence of administrative fragmentation I operationalized it through the local political context generated at the level of city’s subdivisions following local elections results. In this regard I used as case study the electoral behaviors in the six subdivisions of Bucharest, the capital of Romania, at the national legislative elections from 2000-2016.
For the data analysis I employed methodology and statistical tools drawn from the subdiscipline of the electoral geography in order to capture and understand the spatialization of the vote and the concentration of the electoral behaviors in the subdivisions of Bucharest. The aim of the empirical analysis was to see if the local political context in subdivisions develops into an electoral competition that mobilizes and realigns the electorate by disregarding socio-spatial conflicts. Bucharest is divided into six subdivisions (named with numbers from 1 to 6) which have the shape of a circular sector and display mostly similar socio-spatial structures: a central gentrified and richer area and outside of it a large area of socialist collective housing estates. The sectors display similarities because their boundaries were draw during the totalitarian socialist era as an arbitrary mean to achieve socio-spatial standardization within the city. An exception is made here by sector 1 which has large clusters of rich and low-density area and sector 5 which has a larger area of poor and vulnerable groups. The results of the study showed the capacity of Bucharest’s subdivisions to influence electoral behaviors in national legislative elections. However, the results need to be discussed within the framework of the Romanian political system, because the influence is mediated by several variables that determine the relationship between local political context and electoral behaviors. On these matters I develop the remaining paragraphs of the conclusions.
The analysis made on the electoral behaviors in Bucharest at the legislative elections from 2000-2016 revealed that there were elections in which the local political context in sectors matters and the performances are concentrated in certain sectors, yet there are also elections in which the electoral behaviors do not cluster significantly in subdivisions. The strongest clusterization was at the 2008 elections, when the Bucharest subdivisions acted as strongholds for politicians and parties that dominated local politics. At that time most of mayors in subdivisions were serving their second or third consecutive term and consolidated over time their relationship with the local electorate. This clusterization was present at all three actors at the 2008 elections, namely the PDL, the PNL and the PSD + PC alliance, as each dominated the local political contexts and concentrated their best scores in at least one sector. At this election, the influence was also enhanced by the low degree of polarization and ideological substance at the national scale of politics and the extremely low turnout.
In the other elections, the local political contexts had mainly a weak or insignificant influence on the electoral behaviors in subdivisions because the vast majority of mayors were at that time serving for their first term, thus without the time needed to develop strong local loyalties. However, there are also mayors that won a second term, yet without generating a clusterization of electoral sympathies in their subdivisions. I suggested in the analysis that the weak or lack of clusterization at other elections, in this case 2000, 2012 and 2016, may have been also determined by the landslide victory scored by alliances or parties at the local elections. These results created different political contexts at the level of Bucharest in comparison with the 2008 electoral year, when there was no dominant actor and the city was also highly fragmentated from political partisanships. The administrative fragmentation of Bucharest influenced the local electoral behaviors in the city when there was also a political fragmentation of the city.
The analysis of the case study on the electoral behaviors in the six administrative subdivisions of Bucharest revealed a scenario where the socio-spatial conflicts can lose political representation in an administrative and political partitioned large city. This line of research should also be studied in other urban contexts in Europe or on other continents. With this contribution, I hope more attention will be engaged in the study of the relationship between the administrative subdivisions, the electoral behaviors and the socio-spatiality of cities. Many of the major contemporary cities still maintain decades-old administrative partitions that do not respect local socio-spatial conflicts and do not offer them the needed democratic political representation.
Agnew, J. (1990). From political methodology to geographical social theory? A critical review of electoral geography, 1960-1987. In: SHELLEY, F., JOHNSTON, R., TAYLOR, P. (eds.): Developments in Electoral Geography. Londra: Routledge 15-21.
Agnew, J. (1987). Place and Politics: The Geographical Mediation of State and Society. Boston: Unwin.
Agnew, J. (1996). Mapping politics: how contexts counts in electoral geography. Political Geography, 15, 2, 129-146.
Anselin, L. (1988). Spatial Econometrics: Methods and Models. Dordrecht: Springer.
Anselin, L. (1995). Local Indicators of Spatial Association—LISA. Geographical Analysis, 27, 2, 93-115.
Armaș, I., Gavriș, A. (2016). Census-based Social Vulnerability Assessment for Bucharest. Procedia Environmental Sciences, 32, 138-146.
Brenner, N., Theodore, N. (2002). Cities and the Geographies of “Actually Existing Neoliberalism”, Antipode, 34, 3, 349-379.
Buser, M. (2014). Democratic Accountability and Metropolitan Governance: The Case of South Hampshire, UK. Urban Studies, 51, 11, 2336-2353.
Cavalcanti, M. (1997). Urban reconstructions and autocratic regimes: Ceausescu’s Bucharest in its historic context. Planning Perspectives, 12, 1, 71-109.
Chiru, M., Gherghina, S. (2012). When voter loyalty fails: Party performance and corruption in Bulgaria and Romania. European Political Science Review, 4, 1, 29-49.
Evans, G. (2017). Social class and Voting. In: Arzheimer, K., Evans, G.: The Sage Handbook of Political Behavior. London: Sage, 177-198.
Frisken, F., Norris, D. (2001). Regionalism Reconsidered. Journal of Urban Affairs, 23, 5, 467-478.
Getis, A. (2010). Spatial Autocorrelation. In: Fischer, M., Getis, A. (eds): Handbook of Applied Spatial Analysis. Springer, Berlin.
Gheorghita, A. (2014). Determining the Components of Leader Effects in a Post-Communist Context. In: Costa Lobo, M.C., Curtice, J. (eds.): Personality Politics? The Role of Leader Evaluations in Democratic Elections. Oxford: Oxford UP, 191-214.
Gherghina, S. (2013). Going for a Safe Vote: Electoral Bribes in Post-Communist Romania. Journal of Contemporary Central and Eastern Europe, 21, 2-3, 143-164.
Gherghina, S. (2015). Party Organization and Electoral Volatility in Central and Eastern Europe: Enhancing voter loyalty. New York: Routledge.
Gherghina, S., CHIRU, M. (2013). Taking the Short Route: Political Parties, Funding Regulations, and State Resources in Romania. East European Politics and Societies, 27, 1, 108–128.
Gherghina, S., Jiglau, G. (2011). The Ideological Institutionalization of the Romanian Party System. Romanian Journal of Political Science, 11, 1, 71-90.
Gherghina, S., Volintiru, C. (2017). A new model of clientelism: Political parties, public resources, and private contributors. European Political Science Review, 9, 1, 115-137.
Guțoiu, G. (2018). Electoral geography and the post-socialist city: Is the pre-world war II urban tissue still important? Evidence from Bucharest at the Romanian parliamentary elections of 2016. European Quarterly of Political Attitudes and Mentalities, 7, 3, 47-60.
Guțoiu, G. (2019). The Urban Electoral Geographies of Cluj-Napoca, 1992-2016. Online Journal Modelling the New Europe, 29, 108-135.
Johnston, R., Pattie, C. (2006). Putting Voters in Their Place: Geography And Elections in Great Britain. Oxford: Oxford UP.
Kubler, D. (2012). Governing the Metropolis: Towards Kinder, Gentler Democracies. European Political Science, 11, 3, 430-445.
Lowery, D., Lyons, W. (1989). The Impact of Jurisdictional Boundaries: An Individiual-Level Test of The Tiebout Model. The Journal of Politics, 51,4, 73-97
Leib, J., Quinton, N. (2011). On the shores of the "moribund backwater"?: Trends in electoral geography research since 1990. In: Warf, B., Leib, B. (eds.): Revitalizing Electoral Geography. Farnham: Ashgate, 9-37.
Manzetti, L., Wilson, C. (2007). Why Do Corrupt Governments Maintain Public Support? Comparative Political Studies. 40, 8, 949-970.
Marcińczak, S., Gentile, M., Rufat, S., Chelcea, L. (2014). Urban Geographies of Hesitant Transition: Tracing Socioeconomic Segregation in Post‐Ceauşescu Bucharest. International Journal of Urban and Regional Research, 38, 4, 1399-1417.
Maxim, J. (2019). The Socialist Life of Modern Architecture: Bucharest, 1949-1964. New York: Routledge.
O'Loughlin, J., Flint, C., Anselin, L. (1994). 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, 351-380.
O'Sullivan, D., Unwin, D. (2003). Geographic Information Analysis. Hoboken: John Wiley Sons.
Pattie, C., Johnston, R. (2000). People who talk together vote together: an exploration of contextual effects in Great Britain. Annals of the Association of American Geographers, 90, 1, 41-66.
Rufat, S., Suditu, B. (2008). Bucarest, faire la ville entre héritages et mimétisme. L’Espace géographique, 37, 4, 327-337.
Shin, M., Agnew, J. (2008). Berlusconi’s Italy. Mapping Contemporary Italian Politics. Philadelphia: Temple UP.
Shin, M., Agnew, J. (2011). Spatial Regression for Electoral Studies. In: Warf, B., Leib, J. (eds.): Revitalizing Electoral Geography, Farnham: Ashgate, 59-74.
Soare, S., Popescu, M., Gherghina, S., Chiru, M. (2013). Stable or Fluid? Making Sense of Party System Changes in Post-Communist Romania. Paper prepared for the 41st ECPR Joint Sessions of Workshops, Mainz. https://ssrn.com/abstract=2635965
Tufiș, C. (2008). Institutional Trust – Victim of the Postcommunist Transition. In: Voicu, B., Voicu, M. (eds.) The Values of Romanian: 1993 – 2006. Iasi, Institutul European, 115-143.
Van Der Wusten, H., Mamadouh, V. (2014). It is the context, stupid!. Or is it? British-American contributions to electoral geography since the 1960s. L’espace politique, 23, 2, https://journals.openedition.org/espacepolitique/3048.
Zarecor, K. E. (2018). What Was So Socialist about the Socialist City? Second World Urbanity in Europe. Journal of Urban History, 44, 1, 95–117.
Zimmermann, K. (2014). Democratic metropolitan governance: experiences in five German metropolitan regions. Urban Research & Practice, 7, 2, 182-199.
Acronyms and names for parties and alliances mentioned in the paper.
LISA maps for parties and alliances of interest in Bucharest at legislative elections 2000-2016
Cite this paper:
Guțoiu, G. (2020), Urban Administrative Fragmentation and Its Consequences on Electoral Behaviors at Romanian National Legislative Elections. A Case Study on Bucharest, European Quarterly of Political Attitudes and Mentalities ISSN 2285-4916 ISSN-L 2285-4916, pp. 14-37.