Birds of a feather flock together : differences in descriptive representation in municipality councils as a result of social and cultural bubbles

(1)

Birds of a feather flock together

Differences in descriptive representation in municipality councils as a result of social and cultural bubbles

Bachelor’s Thesis Sebastiaan Beaujon sebastiaanbeau@gmail.com

10563830

First reader: Seiki Tanaka Second reader: Sijeong Lim

7143 words

Faculty of social and behavior sciences Department of political science

Amsterdam June, 2018

(2)

Introduction

Male dominated politics is a thing of the past. The last decades, more and more people call for more diversification in politics and these calls are increasingly met (OECD, 2018). Nowadays, advocates are increasing in the political landscape and as a result politics has become more diverse. In 2015, the Canadian Prime Minister Trudeau presented its newly formed cabinet that consisted of an equal amount of men and women and one that was ethnically diverse, mirroring the overall Canadian population. When Trudeau was asked why he choose to form a gender balanced and ethnically diverse cabinet, he simply replied: “Because it’s 2015” (Murphy, 2015). This simple but powerful statement might have inspired other political leaders in Western democracies to do the same. In 2017, newly elected French president Emmanuel Macron presented a gender-balanced cabinet and in Spain, Prime Minister Pedro Sanchez took it a step further and appointed 6 men and 11 women (Corbet, 2017; Durantez, 2018). These developments have been applauded by many. Political leaders who do not take gender balance seriously are often criticized. David Cameron, former British Prime Minister, was said to have a ‘women problem’ when he failed to fulfill his promise of giving a third of all minister posts to women (Newman, 2014). The diversification in terms of gender is however still limited to some countries. In the Netherlands for example, women still only hold 36 percent of seats in the House of Representatives and 35 percent in the Senate (Eerste Kamer, n.d.; Parlement & Politiek, n.d.). Although these numbers are relatively high compared to other countries, this is still far from ideal (OECD, 2018). Besides the lack of women in democratic political institutions, members of minority groups in general lack descriptive representation. Democratic political institutions are still relatively uncolored and consist mostly of white, middle-aged individuals. The new American Senate is the most racial and ethnically diverse in American History but members of racial or ethnic minorities still only make up 19 percent (Krogstad & Bialik, 2017). Although progress has been made, there is still a large and undesirable gap. This gab is even larger in the Netherlands, where the current House of Representatives counts zero black representatives (Bahara, 2017).

These same discrepancies can also be seen on the Dutch local levels. As the data of this research will show, both the municipality council of Rotterdam and The Hague consist mostly of white, middle-aged and highly educated people and in the case of Rotterdam, white, middle-aged and highly educated men. This is troubling because the underrepresentation of these groups in politics might for example lead to alienation of these groups and endanger the stability of democracy. Although nationally, members of racial and ethnic minority groups make up less a proportion then autochthones, the underrepresentation of minority group members can be a result of the overrepresentation of autochthonous people in the population of rural areas (McNamara 2017, p. 25). As this research will show, in larger municipalities like The Hague and Rotterdam, the amount autochthones and members of racial or ethnic minority groups are similar.

(4)

An often-opted reason behind the lack of descriptive representation in democratic political institutions is a low voter-turnout. White, middle-aged and highly educated people tent to vote more than people with other descriptive, like minority group members are the lower educated. When turnout is low it are mostly white, middle-aged and highly educated people that vote. As a result, this leads to an heterogenous municipality council. It seems however that turnout and demographics can not fully explain the variations in de descriptive representation. Although the municipality of Rotterdam and The Hague share equal demographic compositions and levels of voter-turnout, the descriptive characteristics of the municipality councils differ.

The academic field concerning descriptive representation is a broad one and has experienced an increase in attention, not only in the academic world. More and more countries implement minority or women quotas to stimulate the descriptive representation of these groups (McCann, 2013). As stated earlier however, we are far from an ideal situation in the Netherlands. This is not necessarily due to a lack of knowledge because also in the Netherlands this has been thoroughly researched. This research however, is mostly conducted on a national level. This is not necessarily strange, people perceive national politics to be more important than local politics (Boogers, Van Ostaaijen & Slagter, 2010, p. 22; Reif & Schmitt, 1980, pp. 8-10), but as a result there is still much to learn about the descriptive representation in a lower political level. The research that has been done focusses mostly on the municipality level itself, using broadly excepted theories. However, as noted earlier, these cannot explain all the differences in regard to descriptive representation since the demographic composition and levels of voter-turnout in the municipalities of Rotterdam and The Hague are similar.

Most of the research mentioned above uses data on the municipality level to come to certain conclusions. As the data of this research will show, the overall demographics of both municipalities are similar but on a smaller scale the demographic compositions differ. Breaking down the municipality data and look for differences on a lower level, like on a district level, might show different results and could explain the variation in descriptive representation of the municipality councils of Rotterdam and The Hague.

Why this difference is present, will be the main question of this thesis. The structure of the thesis is as follows. First, I will briefly state the importance of this research. The second part will examine the main theories regarding the issue and importance of descriptive representation and give possible theoretical explanations for the difference in descriptive representation whilst voter-turnout and demographic composition are equal. Third, the research method will be laid down, explaining how the research is done and which data has been used. In the fourth part, the research will be conducted and the research question will be answered. This this will end with a concluding chapter.

(5)

As many authors have shown (Rosenstone & Hansen, 1993; Verba, Schlozman & Brady, 1995), whether national and local elected officials represent the population of a municipality depends on some degree on the voting turnout during elections, because some groups are more likely to vote then others. When voting-turnout is low, the composition of the people that vote does not represent the entire population that is allowed to vote. As a result, the elected officials do not proportionally represent the people they are supposed to govern (Wattenberg, 1998, pp. 11-14). There are however also differences in representation between municipalities that show similar levels of voter-turnout. In this case, the difference in representation cannot be explained by a difference in voter-turnout and other mechanisms play a part. In order explore these relatively unexplored grounds it is important to lay down a solid theoretical foundation, which will be done in the following section. First, the difference in turnout between the Dutch national elections and municipality elections will be examined. This will create a better understanding of why turnout during the municipality elections is low, which is essential for the rest of this thesis. Secondly, the implications of a low turnout will be discussed, and especially the relation between low voter-turnout and representation. It is important to discuss this because, although the voter-voter-turnout of the municipalities that will be examined are similar, they are low. Representation is a broad concept that knows many variations, therefore, in the third part I will distinguish the two forms that I believe are most important, namely substantive and descriptive. Descriptive representation can to some degree result in substantive representation too and will therefore receive the main focus in this thesis. After diving deeper into the two different forms of representation and their mutual relation, the literary overview will be concluded by examining the possible reasons why representation might differ despite similar voting-turnout levels and demographic compositions.

National voting-turnout versus municipality voting-turnout

The last decades, many Western European countries, including the Netherlands, have experienced decentralization of government tasks and territorial reorganizations, slowly giving local governments more decisional power and allowing them to act more autonomously (Van Nijendaal, 2014, pp. 85-87; Morlan, 1984, p. 457). The main reason behind these movements was, and still is, to increase the quality of public services and the efficiency of the local governments. People hoped that the increase in decisional power over existing domains and the increase in domains that municipalities are able to shape would also positively influence the way its citizens would look towards the municipality. Because the decisional power of municipalities increased, citizens would believe them to be more important. In turn this should have led to an increase in political participation (Morlan, 1984, p. 457). This is not a far-fetched thought, but the latter unfortunately never happened. The voting-turnout during municipality elections has steadily decreased from 74 percent in 1978 to 55 percent in 2018. During that same period, the voter-turnout during the general elections has decreased as well but has also seen a revival, resulting in a turnout that has been fluctuating around 80 percent the last 15 years (I&O Research, 2017; Kiesraad, 2018). Although general interest in politics can not solely be measured by voting turnout, this data is a

(6)

good indication that the general interest in national politics has remained relatively high over time, with some larger fluctuations in the 90’s, while the general interest in municipality elections has steadily decreased.

“[T]he general notion [is] that a person will vote if the information and time costs of doing so are outweighed by the benefits of potentially casting the deciding vote and the rewards (or avoided costs) from voting” (Harder & Krosnick, 2008, p. 541). This notion will be broken down in order to fully understand the difference in turnout between the so-called first- and second-order elections, national elections and local, regional and supranational election respectively. Reif and Schmitt (1980) developed the second-order framework, with which they try to explain this increasingly relative phenomenon. They start with the notion that there exists a hierarchy between the two elections, with first-order elections being more important than second-order elections, because, as they state, second-order politics are mainly determined by national politics. When it comes to the policies that are produced, voters therefore tend to perceive the policies produced by municipality councils as less important than policies produced at the national level (Boogers, Van Ostaaijen & Slagter, 2010, p. 22; Reif & Schmitt, 1980, pp. 8-10). Although the information and time costs of voting remain the same, the benefits of potentially casting the deciding vote and the rewards from voting are less than in national elections. This could partly explain the lower voter-turnout. However, as noted before, voter-turnout during municipality elections is not only low, it has been decreasing the last 40 years. We have to be critical when it comes to the above-mentioned statement because, although turnout has decreased the last decades, the importance of municipalities and their decisional power has increased during that same period. In the Netherlands, since the beginning of the decentralization movements, municipalities have received increasing decisional power over local educational, cultural and recreational affairs, accommodation, transportation and order and safety (Allers, Steiner, Hoeben & Geertsema, 2013, pp. 21-22). Since 2015, municipalities are also responsible for youth care, labor participation, social support and partly healthcare (Pommer & Van der Torre, 2014, pp. 69-70; Van Nijendaal, 2014, pp. 86). This last development especially has increased the importance of municipality councils, which might influence the perception of voters and in turn increase the turnout during future municipality elections. The turnout during the municipality elections of 2018 was a few percent points higher than the previous election of 2014, but we have to wait for future municipality elections in order to conclude whether this is a development that sparks a revival of municipality turnout or if it is just another fluctuation. Reif and Schmitt could not have included this, since the decentralization in most Western European countries occurred after they published their article and framework, but nevertheless, the theory they created is still valid, because, as Boogers, Van Ostaaijen and Slagter (2010, p. 22) note, around 80 percent of the dutch population believes that the municipality councils have little to no influence on their daily lives.

On top of that, various other reasons influence the perception of voters regarding the importance of municipalities. Not only voters, but also political parties and the media often see order elections as less important. Political parties reserve less financial funds for

(7)

second-order campaigns (Put, Maddens & Smulders, 2015, p. 139), commit less energy to these elections and top politicians often do not run for municipality council seats. This could discourage mobilization and effect the perception of voters as well. If the political parties seem less interested in local politics, why should voters be? It potentially works the other way around as well, when voters aren’t as interested in local politics as they are in national politics, political parties might focus more on national elections and less on local ones (Lefevere & Van Aelst, 2014, pp. 167). The perception of voters and political parties regarding the importance of municipality elections also influences the attention media devotes on covering them and the corresponding campaigns. If voters care less about local elections and political parties devote less attention on them, media will lack incentives to cover the elections and campaigns (De Vreese, 2003; Lefevere & Van Aelst, 2014, p. 162). This again, works the other way around as well. If media do not provide the space for political parties to vent their positions and make themselves known, it becomes harder for political parties to inform voters. It becomes harder for voters to inform themselves and this can decrease voter-turnout because uninformed voters are less likely to vote (Lassen, 2005, p. 110). It almost seems that we are in a vicious circle, where each’s disinterest in local politics increases the disinterest of the other parties.

To conclude this part, lets refer back to “the general notion that a person will vote if the information and time costs of doing so are outweighed by the benefits of potentially casting the deciding vote and the rewards (or avoided costs) from voting” (Harder & Krosnick, 2008, p. 541). As stated earlier, the benefits of potentially casting the deciding vote and the rewards from voting are less than in national elections. On top of that, because both political parties and media do not pay much attention to local elections, it becomes harder for voters to acquire information. Therefore, the information and time costs a person has to pay are a lot higher than when national elections take place. The information and time costs do not outweigh the benefits of potentially casting the deciding vote and the rewards from voting. This explains why the number of voters that turn out during municipality elections is low, and a lot lower than during national elections.

The Implications of a low voter-turnout

If everyone is the same, no one will turn out to vote because to costs of voting outweigh the benefits gained from voting. However, the information and time costs a potential voter has to pay in order to vote are relative. The physical costs of voting differ per voter but are dispersed equally among social and economic groups and therefore do not affect the (in)equality of voter-turnout. The mental and cognitive costs however are very much in play here. For some voters it is relatively easy to inform themselves about upcoming elections and for others it is not. Some people run into difficulties when collecting and processing information about politics more often than others and others are simply less interested in politics or knowledgeable then others. This creates political participation gaps (Gallego 2010, p. 241). These barriers are easier to overcome if one has many cognitive resources. The main and most obvious way to increase this, is through education. The relation between education and one’s political knowledge and interests has been extensively

(8)

researched by scholars and most of them come to the same conclusion, namely that the higher someone’s education, the more politically informed that person is likely to be (Berelson & Steiner, 1964, p. 423; Lassen, 2005, p. 114; Linder, 2010, p. 112-113; Lipset, 1960, p. 182). Because people differ in educational levels and therefore differ in the political knowledge they possess, the composition of voter-turnout is unequal. As states earlier, uninformed citizens are less likely to vote, therefore people with higher forms of education are more likely to vote than citizens that do not have such an education level. Although this relation does not hold in many Western European countries because strong left-wing parties successfully mobilize the lower educated citizens, Gallego (2010, pp. 242-243) shows that this relation is present in the Netherlands. As a result, when turnout is low, the people that do turn out to vote tend to be the higher educated. And it is not just the higher educated people that are over-represented in voter-turnout, socially privileged citizens in general are more likely to vote than the socioeconomically disadvantaged. Citizens who belong to a racial or ethnic minority or for example have a low income vote less regularly. Age also affects someone’s likeliness to vote. Younger people tend to vote less than older citizens, and the elderly tend to vote less than middle-aged citizens. (Berelson & Steiner, 1964, p. 423; Linder, 2010, p. 112-113; Lipset, 1960, p. 182; Verba, Schlozman, and Brady 1995, pp. 1-2). As a result, people that vote are more likely to be white, middle-aged, high educated people with an income that is above average. This creates a biased turnout and as turnout decreases, this bias increases (Rosenstone & Hansen, 1993, p. 238; Wattenberg, 1998, p. 15).

This biased, low voter-turnout can cause several issues for the modern-day democracy. The most important issue is the issue of representation, or in this case, the lack of representation. Not only does a low voter-turnout result in a voter-turnout that does not represent the characteristics found of the whole society equally, “who votes, and who doesn't, has important consequences for who gets elected and for the content of public policies” (Lijphart, 1997, p. 4). As the research of Griffin and Newman (2005) indicates, the preferences and ideals of voters are better represented in policies then those of non-voters, a “government responds disproportionately to the politically active” (Griffin & Newman, 2005, p. 1218). This is not strange, elected officials focus mostly on voters and their preferences because they are the ones that voted the officials into office. If an elected official runs again, it is up to the voters to determine his or her faith (Griffin & Newman, 2005, pp. 1221-1222). Because elected officials are more responsive to voters than to non-voters, the policies these officials enact are often biased against non-voters. Because non-voters have different characteristics then voters, the policies enacted favor specific socioeconomic groups and disdain the disadvantaged, i.e., the less educated, the young and elderly, and racial and ethnic minorities. As Lijphart (1997, p. 1) accurately phrases: “Unequal participation spells unequal influence”.

Representation

Representation itself is not a contested concept, but what should be represented is up for debate. When speaking about representation in a political sense, two important forms can be

(9)

distinguished, namely substantive representation and descriptive representation. Substantive representation entails that the elected officials should mirror the interest and preferences of the electorate (Mansbridge, 1999, p. 630). Descriptive representation entails that a representative represents the ascriptive characteristics that are politically relevant, like gender, ethnicity, age and education of its constituents (Mansbridge, 1999, p. 629; Pantoja & Segura, 2003, p. 443). If a municipality council is only filled with white, highly educated, middle-aged people that promote and protect the interests of all socioeconomic groups in society, including low educated members of an ethnic or racial minority group, a substantive representation advocate would consider this council as representative. It is of course very unlikely that a municipality council is only filled with white, highly educated, middle-aged people but I believe it is also very unlikely that a council that is filled with white, highly educated, middle-aged people promotes and protects the interests of all socioeconomic groups proportionately. I’m not saying that a white, highly educated, middle-aged person cannot protect and promote the interest of a person from a different socioeconomic group, but I believe there has to be some degree of descriptive representation in order to serve and promote the interest of all socioeconomic groups. On top of that, each of these characteristics influence one’s interest and preferences and therefore also result, to some extent, in substantive representation.

Many scholars have researched the gender gaps in political attitude and have come to the conclusion that men and women prioritize different topics and differ in their political attitudes. Women tend to be more supportive towards liberal policies and candidates than men and men tend to be more supportive towards conservative policies and candidates than women (Norrander, 2008). The political attitudes of men and women differ because they tend to hold different social roles. These social roles influence one’s socialization process and experiences, “lead[ing] to different interpersonal and intrapersonal processes that in turn elicit different political attitudes for men and women” (Diekman & Schneider, 2010, p. 489). These differences in attitudes include that women favor communion more than agency and men vice versa. As a result, women show more support for policies that lessen hierarchical structures, such as policies that support disadvantaged groups, and hold more progressive racial attitudes than men. Men, on the other hand, tend to favor policies that support the status quo, maintain existing hierarchies and promote competition (Diekman & Schneider, 2010, p. 490; Schmitt & Wirth, 2009, pp 432-433). But it is not just gender that creates different political attitudes. Research has shown that ethnic minorities and majorities differ in political attitudes too (Staerklé, Sidanius, Green & Molina, 2010, pp. 492-493, 512). Also, one’s age affects one’s political attitudes and values. As Inglehart (1977, pp. 52-53) notes, younger people tend to be more post-materialistic, older people tend to be more materialistic and when age progresses, one becomes more materialistic and less post-materialistic. These differences have political implications. “Materialist values reflect a relatively strong attachment to maintaining order and preserving economic gains. Post-Materialist values emphasize individual self-expression and achieving a more participant, less hierarchical society” (Inglehart, 1977, p. 179). One’s education influences one’s political attitudes and values as well.

(10)

Someone’s personality is the product of experiences, which include educational experiences. Different forms of education create different personalities (Gerber, et. al., 2010, p. 111). Education influences which values an individual prioritize because “the primary function of education is to expand one's intellectual horizons through, among other things, exposure to a great heterogeneity of people and ideas” (Golebiowska, 1995, p. 24). In turn, an increase in education level increases political tolerance (Golebiowska, 1995, p. 27), and often result in holding more liberal social policy views (Gerber, et. al., 2010, p. 121). Therefore, individuals with high levels of education are for example more likely to support immigration then people that are less educated (Hainmueller & Hiscox, 2007, pp. 436-437). Because individuals with different ascriptive characteristics tend to hold different political attitudes and prioritize different values, it is important that a municipality council is descriptive representative. This is however not the only argument.

As Mansbridge (1999, p. 644) notes, “When interests are uncrystallized, the best way to have one's most important substantive interests represented is often to choose a representative whose descriptive characteristics match one's own”. It is often hard to predict which issues surface during a deliberation process. Therefore, based on descriptive similarities, voters can expect that the way a representative acts is akin to the way the voters would have behaved (Mansbridge, 1999, pp. 636, 644). During deliberation or conflicts it is more than fair that the parties concerned are represented in the assembly by proportionate numbers that reflects their numbers in society, so that every group has a proportionate say in the matters that concern them as individuals as well as a group (Mansbridge, 1999, 634). The range of views that are vivid in society are hard to represent by an assembly with similar characteristics, therefore, having a diverse assembly strengthens the deliberation process and policies that are produced (Dovi, 2002, p. 730), but also increase the legitimacy of the policies and the political institutions that created them. To ensure the stability and legitimacy of a democracy, its citizens should value democratic principles and feel a connection with those who exercise political power (Pantoja & Segura, 2003, p. 441). Not having one’s gender, ethnicity or for example age represented in the governing assembly can decrease this connection, resulting in political alienation. Political alienation in turn results in decreasing trust in political institutions, participation and the feeling of political efficacy. When descriptive representation increases, trust in political institutions, participation and the feeling of political efficacy increases too (Pantoja & Segura, 2003, p. 443; Banducci, Donovan & Karp, 2004, pp. 538-539). “Awareness that one's interests are being represented with sensitivity, and knowledge that certain features of one's identity do not mark one as less able to govern all contribute to making one feel more included in the polity. This feeling of inclusion in turn makes the polity democratically more legitimate in one's eyes” (Mansbridge, 1999, p. 651). Because of all of the above I believe it is very important to look at descriptive representation of municipality councils and therefore, this thesis will focus on this form representation.

Although the above mentioned explains why descriptive representation is important, it is not yet clear show how, despite equal levels of voter turnout and overall demographics, descriptive representation can differ between the municipality councils. Leaning on previously conducted

(11)

research regarding this subject is difficult simply there is not much available. A possible reason for this is that academics often neglect the differences on lower levels and instead use data on a municipality level to explain and analyze descriptive representation. There are however theories from other sociological sciences that might provide an explanation. The main theory, which will receive the most attention here, focusses on bubbles.

Bubbles

The above mentioned descriptive characteristics, ethnicity, age, education and gender, define one as a person and at the same time make one a member of a group that is characterized by that specific characteristics, for example the highly educated. Individuals tend to interact more with people with a same level of education. Although this this not hold when it comes to gender, people with racial and ethnic similarities tend to interact more with people that share those characteristics than with people that do not. As the saying goes, birds of a feather flock together. When people interact mostly with members of similar classes, as one might call them, one neglects, unknowingly, the experiences, preferences and views of people that do not belong to that group. People surround themselves with a bubble and within that bubble, like-minded people are concentrated. This does not only apply on some distant and narrow-minded individuals, “nearly all of us live in cultural and political bubbles” (Hochschild, 2017, p. 422). This does not only change the way people see the world, is also affects their behavior when encountering people from other groups (Clair, 2015, p. 20). Living in such bubbles could create political polarization. A form of such political polarization is the deep cleavage between people who support establishment politicians and people that oppose them, often referring to them as distant elites, and accuse established politicians and parties of failing to account for those that do not belong to the same classes (Mair, 2013, p. 1). This is a well-known and increasingly relevant example to explain the support for protest-parties, but this also shows the social, cultural and political limitations that result from polarization because the bubble one is in, blocks alternative views and ways of life. The lack of interaction between different social, economic or cultural classes reproduce and strengthen the actions and preferences that distinguish one group from another and “differences [are] reinforced by the environment you are in and the everyday interactions you have” (McNamara, 2017, pp. 25-26). Although interacting with others will not pop the symbolic bubble, it might broaden one’s mind and increase affection with those of other social, cultural and economic classes. One is then more able to see themselves in the shoes of others and sympathize with the issues and concerns of others, like political alienation or underrepresentation. A bubble in not only defined by its own members, but also by what lies beyond (McNamara, 2017, p. 24-26). Being in a bubble influences not only one’s actions but also one’s thoughts because people tend to focus on those issues that concern them directly. “Through exposure to multiple viewpoints, people become more aware of relevant information and are more able to think through all of the outcomes of a decision, and so societies will make better collective choices on important matters at all levels of government (Munson, Lee, & Resnick, 2013, p. 420).

(12)

People with a more diverse sphere of interaction could share collective emotions with those from other groups. These collective emotions result from being a member of a certain society or group in which the “individuals may experience emotions, not necessarily as a response to their personal life events, but also in reaction to collective or societal experiences in which only a part of the group members have taken part” (Bar-Tal, Halperin & De Rivera, 2007, p. 442). The underrepresentation of disadvantages groups in politics can be such an experience and the collective emotional reaction could be hope, hope that disadvantaged groups and its members will no longer be disadvantaged. In that sense hope could create the aspiration to achieve common goals, like equality and prosperity (Bar-Tal, Halperin & De Rivera, 2007, p. 449). “Once hope arises, it serves as a prism for the worldview as well as a source for collective mobilization and action to achieve the set goal” (idem).

With regard to descriptive representation, one could argue that people that have a more diverse sphere of interaction look beyond their own descriptive characteristics and take the descriptive characteristics of disadvantages groups into account when casting their vote, if share an emotional bond. “A collective emotion may lead to common action with a group goal even though individual members of the group may experience different personal emotions because they occupy different roles in the group” (Bar-Tal, Halperin & De Rivera, 2007, p. 443). The bubble one is in is then not simply defined by a common attribute, like education, but also by emotional attributes. Such a common goal can for example be more women in politics. More people would than consider to vote for a woman instead of men because unequal representation does not benefit society as a whole or because those people affiliate emotionally with the struggles of, in this case, women (Bar-Tal, Halperin & De Rivera, 2007, p 443; McNamara, 2017, p. 26).

Research

design

In order to find if there is some evidence that indicates that the people in either The Hague or Rotterdam live in such bubbles and therefore come less into contact with alternate viewpoints, this thesis will analyze the demographics of each district and compare the corresponding voting behavior of those people. Voting behavior in this case is measured by the number of votes each council member received in a particular district during the municipality elections of 2018. Each council member naturally has certain descriptive characteristics, and this research will show if people with certain descriptive characteristics tend to vote for candidates with similar ones. In the following section certain choices, like case selection, will be examined as well as the research method. Furthermore, the section below will also elaborate on the data gathering process.

Variable selection and measurement

In order to research representation, it is necessary to define what form of representation will be used and which characteristics will be valued in terms of that form of representation. As noted earlier, and substantiated as well, this thesis will focus on descriptive representation. The

(13)

Table 1. Demographic composition of each municipality

descriptive characteristics that will be examined are the ones that academics label as politically relevant, namely gender, ethnicity, age and education (Mansbridge, 1999, p. 629; Pantoja & Segura, 2003, p. 443).

In this thesis, ethnicity will be categorized the same way this is done in both municipalities, where the distinction is made between Western, non-Western ethnicities and autochthones. Non-Western ethnicities are in turn broken down into Turkish, Surinamese, Moroccan and other non-Western ethnicities. This thesis will make that same distinction but will focus mostly non-Western and non-Western ethnicities as a whole. One belongs to one of these groups if either one’s mother or one’s father is born outside of the Netherlands. It is therefore the origin of one’s parents that define one’s ethnicity and not their own place of birth (CBS, 2002).

Age and education are in this thesis categorized by the categories used by the municipalities. In the case education this entails the distinction between low, medium and high levels of education. With regard to the Dutch education system, a low level of education entails completing the first phase of high school, like completing VMBO or the first three years of VWO and HAVO, or elementary school. A medium level of education is considered to be last 3 years of VWO and HAVO or an MBO study. One has a high level of education if he or she has completed an HBO or a university study (CBS, 2014). Age is operationalized by 5 categories, namely 0-14, 15-24, 25-44, 45-65 and 65+. The first one is obviously of no use in this thesis.

Case selection

As noted earlier, the cases that will be examined are Rotterdam and The Hague. This selection is based on two factors. Firstly, an important issue with doing research in general is accounting for external influences, variables that are not included in the research but could influence the examined relation. In this case, as has been laid down in the theoretical overview, voter-turnout and overall demographic composition could influence the representation of municipality councils. If differences are present, it becomes impossible to find a correlation between the presence of bubbles and the descriptive representation of municipality councils. One therefore has to control for those external variables. This thesis has therefore chosen The Hague and Rotterdam as cases because these municipalities have similar levels of voter-turnout and demographic compositions. The turnout in Rotterdam was 47 percent (Onderzoek en Business Intelligence, 2018, p. 11), in The Hague it was 48 percent during the municipality elections of 2018 (Municipality of The Hague, 2018).

When it comes to

demographics, focusing on the above mentioned descriptive characteristics,

(14)

the two are similar as well. Table 1 shows that the distribution of age levels, the ethnic composition, the education levels and gender are similar in both municipalities. Secondly, the available data of municipalities with regard to the municipality elections differs greatly between municipalities. I therefore had to exclude certain municipalities, like Amsterdam, because, although similar in demographic composition and voter-turnout as well, they did not keep track of the votes each candidate got in each district.

Data

Most of the data used in this thesis can be found in the database of the Centraal Bureau voor de

statistiek (CBS). The dataset used in this thesis contains information about almost all Dutch

neighborhoods. This dataset includes data about the distribution of age, gender and ethnicity (CBS, 2017). This does however not include educational levels. The municipality of Rotterdam has this openly available and uses the CBS definitions to categorize its citizens (De Graaf, 2018), and the municipality of The Hague has data available as well but the latter uses a different definition for each category, making it impossible to combine the data. Fortunately, in 2017, various institutions conducted a large-scale research in commission of the municipality of The Hague, examining the illiteracy in each district of The Hague (CINOP Advies, Etil, Kohnstamm, Instituut & Research centrum voor Onderwijs en Arbeidsmarkt, 2017a, 2017b, 2017c, 2017d, 2017e, 2017f, 2017g). Within their research they included a small analysis, studying the different levels of education. These results will be used in this thesis, since their definition is similar as the one used by the municipality of Rotterdam. However, this data might not paint an as accurate picture as the data of Rotterdam because the data of The Hague is mostly based on a study the CBS conducted in 2012, whereas the data of Rotterdam stems from 2015. Although education levels do not vary much over the years (De Graaf, 2018, p. 3), this is important to note. The descriptive characteristics of the candidates are gathered through social media websites, like on LinkedIn, Facebook and Instagram, and the websites of the political parties the candidates belong to (see appendix A and B).

Research method

In this part, the steps taking in the research part will be explained. Firstly, I will shortly show that the descriptive characteristics of the municipality councils of Rotterdam and The Hague vary and to what extent. After that, it will be examined if there is a reason to believe that the previous mentioned bubbles are present and are stronger in one municipality than the other. The third part will analyze these potential bubbles and see if the presence of potential bubbles influences the voting behavior. Potential bubbles will be measured by looking the demographics of each district. In bubbles, like-minded people are concentrated and people that share the same descriptive characteristics tend to be like-minded (see i.e. Diekman & Schneider, 2010; Golebiowska, 1995; Inglehart, 1977; Staerklé, Sidanius, Green & Molina, 2010). The more equal the categories of each

(15)

Table 2. Descriptive characteristics of each municipality council

Table 3. Ethnical composition for each district in The Hague

variable are dispersed within a district, the less likely a strong bubble is present. The population is then more divers and people’s sphere of interaction is as well. If a certain district contains mostly autochthonous, highly educated people, one could suspect that the people there live in a small and strong bubble because there is little interaction with other groups. The people in that bubble are not exposed to different viewpoints and therefore lack the emotional affiliation with other groups. It is then more likely that these people will vote for candidates with similar characteristics.

Analysis

As table 1 shows, the overall demographics of Rotterdam and The Hague do vary not much. There is however one significant difference, namely the percentage of immigrants, which is 6,5 percent higher in The Hague than in Rotterdam. When looking at descriptive characteristics of the municipality councils of both cities, displayed in table 2, it becomes clear that, as theory suggests, both are dominated by white, high educated council members and that more men than women occupy council seats. Between the two municipality councils there is a large difference in gender representation as well. While the amount of men and women in both cities is almost equal, the

municipality council in The Hague counts 5 more women than Rotterdam’s. With regard to education, age and ethnicity, both municipality councils do not seem to differ much. However, as seen in the overall demographics of both municipalities, Rotterdam counts significantly less immigrants then The Hague. The municipality council of Rotterdam however has more immigrants occupying council seats than The Hague’s. Although immigrants are underrepresented in both municipality councils, the council of Rotterdam seems to represent immigrants better than the municipality council of The Hague. As for gender, it is just the other way around. To understand these differences and find a possible explanation, the following section will analyze the ethnical composition of the municipalities on a district level in order to find potential bubbles.

(16)

Table 4. Ethnical composition for each district in Rotterdam

Ethnicity

As table 3 shows, there are large differences in the ethnical layout between each district in The Hague. In the Laak district, 75,2 percent are immigrants while in Loosduinen immigrants make up 29,9 percent of its population. As table 4 shows, large differences can be found in Rotterdam as well, but there are important differences between the two municipalities. In almost all districts of The Hague, except for Haagse Hout, the difference between the number of immigrants and autochthones is around 20 percent or above. However, in Rotterdam, 4 districts experience low differences between the percentage of immigrants and autochthones, namely Rotterdam centrum,

Noord, Karlingen-Crooswijk and IJsselmonde, two of them even below 2 percent. The people that

live in these areas are more likely to have a diverse sphere of interaction because at least 40 percent of the people that surround them have a different ethnicity. People are exposed to a large variety of viewpoints through the interaction with their surroundings and the differences are then not reinforced but broken down (McNamara, 2017, pp. 25-26). This could increase a feeling of collectivity with collective emotions and actions as a result. In turn, the voters within these districts might vote relatively more often on candidates that are part of disadvantaged groups, like women or members of ethnic minorities.

The difference in amount of autochthones and immigrants is not the only ethnical distinction that can be made. In the district Delfshaven in Rotterdam, 30,5 percent of the population is autochthone. The remaining 69,5 percent are immigrants, and of all sorts. 12,9 percent is Moroccan, 13,1 percent is Turkish, 10 percent is Surinamese, 20 percent is from another non-Western countries and 13,4 percent is a Western immigrant. A similar case can be made for the district Feijenoord in Rotterdam and the districts Laak and Centrum in The Hague. These are ethnically diverse districts and one could therefore expect the people in these districts to vote more on members of disadvantaged groups then people in less diverse districts, like Pernis, Hoek van

Holland and Rozenburg in Rotterdam and Loosduinen and Scheveningen in The Hague. As both

figure 1 and figure 2 show, this is indeed the case. Firstly, these figures show that people do tend to vote on candidates that show similar descriptive, at least ethnical and racial, characteristics. Secondly, these figures do show that there are potentially bubbles in both cities, mostly in the districts where autochthones are overrepresented. Thirdly, these figures show that there are

(17)

potentially more bubbles in The Hague then in Rotterdam because Rotterdam counts more diverse districts. Although some districts in Rotterdam are not diverse, these are the smaller districts, entailing only a small proportion of the overall population (CBS, 2017).

Conclusion

The descriptive characteristics of the municipality councils of Rotterdam and The Hague mostly vary when it comes to gender and ethnicity. As theories have shown, these differences might be a result from the absence or presence of so-called bubbles. When people life in these bubbles their views are limited and their sphere of interaction is as well, resulting in a less diverse voting behavior (Bar-Tal, Halperin & De Rivera, 2007, pp. 442-443, 449). As the results of this thesis show, this is indeed the case. The municipality council of Rotterdam is more diverse than the one in The Hague, especially when the overall demographics of both cities are taken into account. The results have shown that the districts in Rotterdam are more diverse than the ones in The Hague, suggesting that people in Rotterdam live less in bubbles than in The Hague.

References

Allers, M. A., Steiner, B., Hoeben, C., & Geertsema, J. B. (2013). Gemeenten in perspectief (COELO-rapport 13-4)[In Dutch]. Retrieved from

https://coelo.nl/images/rapporten/Gemeenten_in_perspectief.pdf

Bahara, H. (2017, April 30). Nul zwarte politici in de nieuwe Tweede Kamer: hoe pijnlijk is dat? [In Dutch]. Volkskrant. Retrieved from https://www.volkskrant.nl/nieuws-achtergrond/nul-zwarte-politici-in-de-nieuwe-tweede-kamer-hoe-pijnlijk-is-dat-~bce21247/

(18)

Banducci, S. A., Donovan, T., & Karp, J. A. (2004). Minority representation, empowerment, and participation. Journal of Politics, 66(2), 534-556.

Berelson, B. & Steiner, G. A. (1964). Human behavior: an inventory of scientific findings. New York: Harcourt, Brace.

Boogers, M., Slagter, L., van Ostaaijen, J. J. C. (2010). Lokale kiezers: lokale keuzes? Onderzoek

naar de achtergronden en de betekenis van het stemgedrag bij de gemeenteraadsverkiezingen van 2010 [In Dutch]. Tilburg: Universiteit van Tilburg.

CBS. (2000). Standaarddefinitie allochtonen [In Dutch]. Retrieved from

https://www.cbs.nl/NR/rdonlyres/26785779-AAFE-4B39-AD07-59F34DCD44C8/0/index1119.pdf

CBS. (2014). Beroepsbevolking; regio’s 1996-2013 [In Dutch]. Retrieved from http://statline.cbs.nl/Statweb/publication/?DM=SLNL&PA=71887NED&D1=22-25&D2=0&D3=0%2C11-13%2C211&D4=a&HDR=T&STB=G1%2CG2%2CG3&VW=T CBS. (2017). Kerncijfers wijken en buurten 2017 [In Dutch]. Retrieved from

https://www.cbs.nl/nl-nl/maatwerk/2017/31/kerncijfers-wijken-en-buurten-2017

CINOP Advies, Etil, Kohnstamm Instituut & Researchcentrum voor Onderwijs en Arbeidsmarkt (2017a) Laaggeletterdheid in Segbroek. Retrieved from

https://denhaag.raadsinformatie.nl/document/5350942/1/RIS294448_bijlage_2_8_Factsheet_r egionale_laaggeletterdheid_-_Segbroek

CINOP Advies, Etil, Kohnstamm Instituut & Researchcentrum voor Onderwijs en Arbeidsmarkt (2017b) Laaggeletterdheid in Scheveningen [In Dutch]. Retrieved from

https://denhaag.raadsinformatie.nl/document/5350941/1/RIS294448_bijlage_2_7_Factsheet_re gionale_laaggeletterdheid_-_Scheveningen

CINOP Advies, Etil, Kohnstamm Instituut & Researchcentrum voor Onderwijs en Arbeidsmarkt (2017c) Laaggeletterdheid in Loosduinen [In Dutch]. Retrieved from

https://denhaag.raadsinformatie.nl/document/5350946/1/RIS294448_bijlage_2_4_Factsheet_r egionale_laaggeletterdheid_-_Loosduinen

CINOP Advies, Etil, Kohnstamm Instituut & Researchcentrum voor Onderwijs en Arbeidsmarkt (2017d) Laaggeletterdheid in Leidschenveen-Ypenburg [In Dutch]. Retrieved from

(19)

https://denhaag.raadsinformatie.nl/document/5350939/1/RIS294448_bijlage_2_6_Factsheet_r egionale_laaggeletterdheid_-_Leidschenveen-Ypenburg

CINOP Advies, Etil, Kohnstamm Instituut & Researchcentrum voor Onderwijs en Arbeidsmarkt (2017e) Laaggeletterdheid in Laak [In Dutch]. Retrieved from

https://denhaag.raadsinformatie.nl/document/5350944/1/RIS294448_bijlage_2_2_Factsheet_r egionale_laaggeletterdheid_-_Laak

CINOP Advies, Etil, Kohnstamm Instituut & Researchcentrum voor Onderwijs en Arbeidsmarkt (2017f) Laaggeletterdheid in Haagse Hout [In Dutch]. Retrieved from

https://denhaag.raadsinformatie.nl/document/5350940/1/RIS294448_bijlage_2_5_Factsheet_r egionale_laaggeletterdheid_-_Haagse_Hout

CINOP Advies, Etil, Kohnstamm Instituut & Researchcentrum voor Onderwijs en Arbeidsmarkt (2017f) Laaggeletterdheid in Escamp [In Dutch]. Retrieved from

https://denhaag.raadsinformatie.nl/document/5350945/1/RIS294448_bijlage_2_3_Factsheet_r egionale_laaggeletterdheid_-_Escamp

CINOP Advies, Etil, Kohnstamm Instituut & Researchcentrum voor Onderwijs en Arbeidsmarkt (2017g) Laaggeletterdheid in Centrum [In Dutch]. Retrieved from

https://denhaag.raadsinformatie.nl/document/5350943/1/RIS294448_bijlage_2_1_Factsheet_re gionale_laaggeletterdheid_-_Centrum

Clair, R. N. (2015). The Stratification of Cultural Networks. Intercultural Communication

Studies, 24(1).

Corbet, S. (2017, May 17). Emmanuel Macron names gender-balanced cabinet from across political spectrum. Independent. Retrieved from

https://www.independent.co.uk/news/world/europe/emmanuel-macron-cabinet-government-france-election-ministers-named-2017-local-vote-a7741151.html

De Vreese, C. H. (2003). Television reporting of second-order elections. Journalism Studies, 4(2), 183-198.

Diekman, A. B., & Schneider, M. C. (2010). A social role theory perspective on gender gaps in political attitudes. Psychology of Women Quarterly, 34(4), 486-497.

Dovi, S. (2002). Preferable descriptive representatives: Will just any woman, black, or Latino do?.

(20)

Durantez, M. G. (2018, June 9). Spain's new cabinet shows that men can be better than women at achieving gender diversity in politics. Independent. Retrieved from

https://www.independent.co.uk/voices/pedro-sanchez

Eerste Kamer. (n.d.). Vrouwen in de Eerste Kamer [In Dutch]. Retrieved from https://www.eerstekamer.nl/begrip/vrouwen_in_de_eerste_kamer

Gallego, A. (2010). Understanding unequal turnout: Education and voting in comparative perspective. Electoral Studies, 29(2), 239-248.

Gerber, A. S., Huber, G. A., Doherty, D., Dowling, C. M., & Ha, S. E. (2010). Personality and political attitudes: Relationships across issue domains and political contexts. American Political

Science Review, 104(1), 111-133.

Golebiowska, E. A. (1995). Individual value priorities, education, and political tolerance. Political

Behavior, 17(1), 23-48.

Hainmueller, J., & Hiscox, M. J. (2007). Educated preferences: Explaining attitudes toward immigration in Europe. International organization, 61(2), 399-442.

Harder, J., & Krosnick, J. A. (2008). Why do people vote? A psychological analysis of the causes of voter turnout. Journal of Social Issues, 64(3), 525-549.

Hochschild, A. R. (2017). A Response to William Davies’ “A Review of Arlie Russell Hochschild’s Strangers in Their Own Land: Anger and Mourning on the American Right (2016: New York: New Press, 351 pp).”. International Journal of Politics, Culture, and Society, 30(4), 421-423.

Inglehart, R. (1977). The silent revolution: Changing values and political styles among Western

publics. Princeton University Press.

I&O Research. (2017, November 19). Gemeenteraadsverkiezingen 2018: hoe de dalende opkomst tegen te gaan? [In Dutch]. Retrieved April 16, 2018, from

https://ioresearch.nl/Home/Nieuws/ArtMID/445/ArticleID/924/Gemeenteraadsverkiezingen-2018-hoe-de-dalende-opkomst-tegen-te-gaan#.WtS_ytNuaYU

Kiesraad. (2018). Gemeenteraad 21 maart 2018 [In Dutch]. Retrieved from https://www.verkiezingsuitslagen.nl/verkiezingen/detail/GR20180321

(21)

Krogstad, J. M., & Bialik, K. (2017, January 24). 115th Congress sets new high for racial, ethnic diversity. Retrieved from http://www.pewresearch.org/fact-tank/2017/01/24/115th-congress-sets-new-high-for-racial-ethnic-diversity/

Lassen, D. D. (2005). The effect of information on voter turnout: Evidence from a natural experiment. American Journal of political science, 49(1), 103-118.

Lefevere, J., & Van Aelst, P. (2014). First-order, second-order or third-rate? A comparison of turnout in European, local and national elections in the Netherlands. Electoral Studies, 35, 159-170.

Lijphart, A. (1997). Unequal participation: Democracy's unresolved dilemma presidential address, American Political Science Association, 1996. American political science review, 91(1), 1-14. Linder, W. (2010). Swiss democracy: Possible solutions to conflict in multicultural societies. New York: St. Martin's Press.

Lipset, S. M. (1960). Political Man: The Social Bases of Politics. Garden City, NY: Doubleday. Mansbridge, J. (1999). Should blacks represent blacks and women represent women? A contingent" yes". The Journal of politics, 61(3), 628-657.

McCann, J. (2013, November 14). Electoral quotas for women: an international overview.

Parliament of Australia, Research Paper Series, 2013-2014. Retrieved from

http://parlinfo.aph.gov.au/parlInfo/download/library/prspub/2840598/upload_binary/2840598 .pdf

McNamara, K. R. (2017). Bringing Class Back In: Cultural Bubbles and American Political Polarization. In Seminar on the State and Capitalism since 1800.

Morlan, R. L. (1984). Municipal vs. national election voter turnout: Europe and the United States.

Political Science Quarterly, 99(3), 457-470.

Municipality of The Hague. (2018, June 8). Verkiezingen gemeenteraad [In Dutch]. Retrieved from https://www.denhaag.nl/nl/bestuur-en-organisatie/verkiezingen/verkiezingen-gemeenteraad.htm Munson, S. A., Lee, S. Y., & Resnick, P. (2013, June). Encouraging Reading of Diverse Political Viewpoints with a Browser Widget. In ICWSM.

(22)

Murphy, J. (2015, November 4). Trudeau gives Canada first cabinet with equal number of men and women. The Guardian. Retrieved from

https://www.theguardian.com/world/2015/nov/04/canada-cabinet-gender-diversity-justin-trudeau

Newman, C. (2014, February 6). David Cameron's 'women problem' is worse than we thought: he's not even aware of it. Telegraph. Retrieved from https://www.telegraph.co.uk/women/womens- politics/10621577/David-Camerons-women-problem-is-worse-than-we-thought-hes-not-even-aware-of-it.html

Norrander, B. (2008). The history of the gender gaps. In L. D. Whitaker (Ed.), Voting the gender

gap (pp. 9–32). Champaign, IL: University of Illinois Press.

OECD (2018). Women in politics (indicator). doi: 10.1787/edc3ff4f-en (Accessed on 18 June 2018). Retrieved from https://data.oecd.org/inequality/women-in-politics.htm

Onderzoek en Business Intelligence. (2018). Gemeenteraadsverkiezingen 2018: Een analyse op

buurtniveau. Retrieved from

https://rotterdam.buurtmonitor.nl/news/Gemeenteraadsverkiezingen-2018-Een-analyse-op-buurtniveau/68

Pantoja, A. D., & Segura, G. M. (2003). Does ethnicity matter? Descriptive representation in legislatures and political alienation among Latinos. Social Science Quarterly, 84(2), 441-460. Parlement & Politiek. (n.d.). Vrouwen in de Tweede Kamer [In Dutch]. Retrieved from

https://www.parlement.com/id/vh8lnhrre0zv/vrouwen_in_de_tweede_kamer Pommer, E. J., Van der Torre, A. G. J. (2014). Gezondheidszorg [In Dutch]. Jaarboek

Overheidsfinanciën, 69-84.

De Graaf, P. (2018). Feitenkaart Opleidingsniveau Rotterdam op gebieds- en buurtniveau 2015 [In Dutch]. Municipality of Rotterdam. Retrieved from

https://rotterdam.buurtmonitor.nl/documents/Onderwijs

Put, G. J., Maddens, B., & Smulders, J. (2015). Buying local votes: The effect of individual

campaign spending under a semi-open PR System in the Belgian local elections. Local Government

(23)

Reif, K., & Schmitt, H. (1980). Nine second‐order national elections–a conceptual framework for the analysis of European Election results. European journal of political research, 8(1), 3-44. Rosenstone, S. J., & Hansen, J. (1993). Mobilization, participation, and democracy in America. Macmillan Publishing Company.

Schmitt, M. T., & Wirth, J. H. (2009). Evidence that gender differences in social dominance orientation result from gendered self-stereotyping and group-interested responses to patriarchy.

Psychology of Women Quarterly, 33(4), 429-436.

Staerklé, C., Sidanius, J., Green, E. G., & Molina, L. E. (2010). Ethnic minority‐majority asymmetry in national attitudes around the world: A multilevel analysis. Political Psychology, 31(4), 491-519. Van Nijendaal, G. A. (2014). Drie decentralisaties in het sociale domein [In Dutch]. Jaarboek

Overheidsfinanciën, 85-100.

Verba, S., Schlozman, K. L., & Brady, H. E. (1995). Voice and equality: Civic voluntarism in

American politics. Harvard: Harvard University Press.

Wattenberg, M. (1998). Turnout decline in the US and other advanced industrial democracies. Center for the Study of Democracy.

(24)

Appendices

(25)

Birds of a feather flock together : differences in descriptive representation in municipality councils as a result of social and cultural bubbles