Opening an Absolute Majority A Typology of Motivations for Opening and Selecting Coalition Partners

The negotiation of a ‘majority pact’,¹x The majority pact includes the indication of the political groups that form the municipal majority, the identity of the mayor, the aldermen and the president of the social action council. Article 1123-1 et seq. of the Code of Local Democracy and Decentralisation (hereafter CDLD). a document marking the starting point of a new municipal coalition in the Walloon Region (Belgium), is not an insignificant act in local political life as it binds the majority partners for six years. It is the result of negotiations between the representatives of various electoral lists. However, not all Walloon municipalities (262 in total) are concerned by the negotiation of a political majority. Indeed, following the municipal elections of 14 October 2018, 189 political groups won an absolute majority. Of these 189 political groups, 167 decided to exercise power alone. Sixty-four percent of Walloon communes are therefore governed by a single formation that won more than half of the seats in the communal council.
Twenty-two of the 189 political groups that won an absolute majority decided not to govern alone but rather to form a coalition by opening to an additional partner.²x The 22 communes are: Andenne, Ans, Aywaille, Bernissart, Binche, Charleroi, Courcelles, Estinnes, Florennes, Grâce-Hollogne, Ham-sur-Heure-Nalinnes, Marche-en-Famenne, Montigny-le-Tilleul, Mouscron, Péruwelz, Rixensart, Seneffe, Silly, Spa, Theux, Thuin and Wavre. At first sight, this opening may seem surprising as the political groups have the majority of seats in the local council. This opening therefore implies a sharing of local power that is not a priori necessary. The choice to open an absolute majority is a very important one, as it conditions the course of the term of office as well as the adoption and content of municipal public policies.
Through an exhaustive analysis, studying the 22 cases concerned by opening an absolute majority, we identify the motivations that impel political representatives to open an absolute political majority in terms of seats to other partners and to choose one partner over another. The research question underlying this article is therefore formulated as follows: why do political groups that hold an absolute majority of seats in the municipal council following the elections of 14 October 2018 open their majority to a minority party, and why do they choose this party?
To answer this question, we first combine our analysis with previous studies of local coalitions and identify the specificity of our research. Second, we detail the characteristics of the 22 communes that were the subject of this study and explain our qualitative methodological choices. Third, we propose an analytical typology of motivations for opening an absolute majority and choice of partner to propose a model of motivations for negotiating oversized coalitions at the local level when the main partner has an absolute majority of seats in the communal council.

• 1 The Formation of a Coalition

  Since the merger of Belgian municipalities in 1977, the number of local coalitions has increased (Dandoy, 2018, p. 513). How are coalition negotiations at the local level envisaged? They have been the subject of a vast array of scientific literature since the 1960s, which include studies on other levels of power. We present, first, the results relating to the negotiation of political majorities, generally, followed by the results specific to the negotiation of local political majorities, taking into account the Walloon context.

  Generally speaking, negotiations aim at the establishment of a notorious – i.e. public and publicised by its protagonists – and lasting alliance between partisan groups to govern an institution, resulting in joint participation in an executive. (Bué & Desage, 2009, p. 10)

  The multiple theories developed on coalition negotiations have a dual purpose: first, to predict the formation of alliances and, second, to understand the decisions and behaviours of political actors. The starting point for these theories is a rational choice approach, which assumes that politicians act rationally by choosing the least resource-intensive and least costly option in terms of its returns (Downs, 1957, p. 137). Theories and analyses of coalition formation have rapidly expanded and enriched, identifying key motivations for coalition formation. Some theories and analyses have focused on ‘process tracing’ of deviant negotiations such as oversized coalitions (Dumont, De Winter & Andeweg, 2011) or have taken into account social psychological dimensions (Andeweg, 2003), enriching the literature on the subject.
  The first motivation is ‘office-seeking’ (Debus, 2008a). Political parties thus integrate an executive to possess power, realised by mandates (Wille & Deschouwer, 2012, p. 25). In this perspective, political parties tend to minimise the number of partners in the majority to obtain more mandates (Bonnet & Schemeil, 1972, p. 270). They associate themselves only with partners that are strictly necessary to reach a majority, while rejecting superfluous partners. In doing so, they create a ‘minimum winning coalition’ (Riker, 1962, p. 32). In the context of this article, the groups studied do not a priori fit into this perspective because the partner who joins an absolute majority is not necessary. One must then consider ‘oversized coalitions’ (Riker, 1962, p. 54; Volden & Carrubba, 2004), where a political party wins an absolute majority and includes an additional partner in its executive. There are several reasons for this choice. A party may seek to preserve power in the longer term and secure its place in a future majority if it loses its absolute majority (Wille & Deschouwer, 2012, p. 112). Oversized coalitions can also be advantageous in periods of crisis (Baron & Diermeier, 2001; Diermeier & Merlo, 2000) or in institutional contexts (Lijphart, 1984; Sjölin, 1993).
  The second motivation under which political actors choose to form a coalition is ‘policy-seeking’ by establishing an alliance that enables them to implement their electoral programme and achieve desired policy outcomes (Axelrod, 1970; Carrubba & Volden, 2000; Crombez, 1996; Strom, 1990). To do so, political parties can subsequently form a ‘minimal connected winning coalition’ (Axelrod, 1970). The latter is, however, debated (Volden & Carrubba, 2004). From the perspective of a minimal connected winning coalition, political parties tend to ally themselves with their neighbours on the political spectrum so as to minimise ideological dissonance and implement their electoral promises with greater ease (Andeweg, 2011; Olislagers & Steyvers, 2015). Research also shows that desired policy outcomes of parties are more likely to be reflected in coalition agreements and to be achieved when these parties have a key role in the coalition (Debus, 2008a). Moreover, median parties are more often included in an oversized coalition (Andeweg, 2011), and the search for more balanced public policies can be pursued through an oversized coalition (Jungar, 2011).
  The third motivation for political parties is to maintain electoral support and even to attract new voters, in other words, ‘vote-seeking’. This maximisation of electoral support can be understood as a means of gaining power and influence over public policy (Downs, 1957, pp. 34-35). Indeed, with a large number of elected representatives, a political party’s bargaining power and its share in the political game are greater (Dodd, 1974, p. 1097), even if it does not necessarily obtain a greater number of positions within a coalition (Debus, 2008a). However, there is a limit to the search for electoral support: political parties tend not to form a coalition with another party that represents the most significant threat in electoral terms (Thrasher, 1999).
  While these three types of motivation lead to different types of coalition formation, it is necessary to consider the multiple constraints that political parties face in forming coalitions. Numerous constraints have been identified in the literature insofar as “the real world of coalitions is one of constraints” (Strom, Budge & Laver, 1994, p. 307). First, the constraints are institutional because there are a host of legal rules regarding the size and composition of executive coalitions, the nomination of coalitions and the conduct of negotiations (Strom et al., 1994). The effect of bicameralism on the formation of oversized coalitions is debated, and while some authors show this effect (Lijphart, 1984; Sjölin, 1993), others have refuted it (Volden & Carrubba, 2004). On the other hand, when qualified majorities are needed to pass certain reforms, oversized coalitions are necessary (Andeweg, 2011). Second, the constraints are contextual because coalitions are formed in political systems with historical, identity-related and symbolic variations in each country (Magre & Pano, 2019; Müller, Bergman & Strom, 2008). Crisis situations can also explain the formation of oversized coalitions, when existing public policies are rather extreme (Baron & Diermeier, 2001), although these results have been refuted (Volden & Carrubba, 2004). Third, the constraints are partisan because political parties may have specific rules for coalition formation (Strom et al., 1994) and because a high level of factionalising within a party negatively affects its likelihood of entering a coalition (Bäck, 2008). In this respect, not all members of political parties necessarily share the same goals (Budge & Laver, 1986): some aim to occupy a position of power, while others want to realise their ideologies or preserve their electoral support (Bué & Desage, 2009). Some new parties may also find it preferable to join an oversized coalition in order to avoid opposition that is not always electorally rewarding (Jungar, 2011). Fourth, coalition agreements are directly constrained by the political preferences of coalition partners (Müller et al., 2008; Müller & Strom, 2008), who seek to secure their political agenda when tensions within a coalition are high (Klüver & Bäck, 2019). These different constraints constitute path dependency factors that shape political coalitions (Müller & Strom, 2008). As a result, the search for coalition partners takes place in the ‘shadow of the past’ (Müller et al., 2008, p. 15). Past experiences are therefore highlighted to predict the composition of governments (Steyvers, Reynaert, De Ceuninck & Valcke, 2008).
  If Belgium is a state affected more than others by oversized coalitions (Andeweg, 2011, p. 197), an analysis of coalition formations in the Belgian political system should ultimately take into account the multiple lessons learned from the scientific literature concerning the congruence of majorities between the different levels of power, given the federal and decentralised institutional organisation (Wille & Deschouwer, 2012, p. 112). Although it has been shown that coalition formations at the regional level are different from those at the national level, in particular in Belgium (Downs, 1998), and that regional political parties may adopt programmatic positions that differ from those of their federal organisation, in particular in Germany (Debus, 2008b), other studies have shown the importance of congruence between levels of power in coalition formations (Bäck, Debus, Müller & Bäck, 2013). For example, congruence is complete when the same parties are members of the executives at the different levels of power (Deschouwer, 2009). Congruence is also sought when there are no national political parties, as is the case in Belgium (Deschouwer, 2009). The main reason political parties seek congruence is the presence of a party at distinct levels of power, which allows for consistency in the way public policies are oriented, with the consequence that the credibility of the political party increases (Deschouwer, 2009).
  While the many studies cited provide a fairly clear picture of the reasons for coalition formation, particularly in the political and legal context of Belgium, it is important to remember that coalition formation at the local level is quite different from that at the national and regional levels (Laver, 1989). For example, a host of differences have been observed, particularly in the UK context, as regards political gains, policy positions and negotiation contexts (Laver, Rallings & Thrasher, 1987). In addition, there are differences in regulatory concerns reflecting a norm of political consensus between actors, especially in the Danish context, for understanding oversized coalitions at the local level (Serritzlew, Skjæveland & Blom-Hansen, 2008). This means that local coalition formations deserve special attention, and this is the justification for this article.
  The negotiations of political majorities at the local level in Belgium, in general, and in Wallonia, in particular, present certain specific characteristics. First, with regard to the theme of ‘minimal connected winning coalitions’, it has already been found that political programmes seem to be less decisive, as differences between parties are not insurmountable at the local level (Wille & Deschouwer, 2012, p. 117).
  Second, the importance of institutional constraints at the local level must be emphasised. Indeed, these constraints weigh particularly heavily on the formation of local coalitions when the legal rules impose the direct election of the mayor, which reveals a form of ‘presidentialisation’ of local politics (Copus, 2019). The mayor is thus a ‘powerful player’ in the game of coalition building (Debus & Gross, 2016; Strom & Swindle, 2002), thereby weakening the influence of local councillors over him (Copus, 2019). It should be noted, however, that the mayors’ confidence in achieving their political projects seems to be lower when they are directly elected (Ervik, 2015). The Walloon Region is particularly concerned by these results since, according to Article L1123-4 CDLD,

  a councillor is elected by right as mayor if he or she is of Belgian nationality and has obtained the most preferential votes on the list that obtained the most votes among the political groups that are parties to the majority covenant.

  Let us emphasise the Walloon specificity, which implies that the negotiation of the local political majority takes place before the mayor is appointed. It is therefore necessary to be a party to the majority pact in order to be appointed mayor. As a result, it may happen that an elected official with the highest number of preference votes is not appointed mayor if his or her party is not included in the majority pact (Grandjean, 2016; Matagne, Radoux & Verjans, 2011).
  Third, in terms of factionalising, the diversity of expectations can lead to dissension within political parties, especially in the case of an absolute majority (Wille & Deschouwer, 2012, p. 104). This dissension sometimes leads a party to form an ‘oversized coalition’ when the group feels weakened by the demands of one or more members or by the difficulties posed by a limited number of local councillors in adopting certain decisions, among other reasons (Laver & Schofield, 1990; Wille & Deschouwer, 2012, p. 22). The political party then allies itself with an additional list to preserve a stable majority throughout the term of office, even though this may lead to conflicts between the running mates (Nasseaux, 2002, p. 6) or a longer decision-making process (Lupia & Strom, 2008, p. 13).
  Fourth, as regards the political preferences of coalition partners, it has been shown that local majority negotiations take into account ‘political history’ (Grandjean, 2016) and that there is a greater probability of existing coalitions renewing than forming new collaborations (Wille & Deschouwer, 2012, p. 25). This is because partners know each other better and know what they can get from each other (Bäck, 2003, p. 463).
  Fifth, with respect to congruence of majorities between the different levels of Belgian government, specificities have been identified with regard to the local level. Thus, when a political party has one or more ministers within the higher levels of power, contacts with the local political majorities to which this party belongs are facilitated (Wille & Deschouwer, 2012, p. 112), allowing certain projects to advance, in particular via more subsidies or information (Ibid., 129). A political party thus opens up its majority to another partner to have additional relays at higher levels (Dumont, De Winter & Ackaert, 2008). It should be noted that congruence between political majorities also allows members of a political party that is active at higher levels of power to influence the negotiations of a municipal political majority (Dumont et al., 2008). Representatives of the higher levels of a party have thus already encouraged a local section to ally itself with a certain partner in a municipal college (Dumont et al., 2008; Grandjean, 2016) or, on the contrary, have shown resistance to a particular coalition, leading, in some cases, to a failure of the negotiations (Dumont et al., 2008).
  Sixth, given the numerous oversized and non-incumbent coalitions at the local level in Flanders, Ellen Olislagers identified the need to distinguish between ‘necessary and supporting mechanisms’ for such coalitions. She demonstrated that “imperfect information, good and bad relations between parties and politicians and higher party involvement stimulated the formation of this kind of coalitions” (Olislagers, 2013, pp. 167-208).
  Compared with the existing literature, we focus on the process of majority formation. Our research is in line with that of authors who are interested in ‘process tracing’ (Dumont et al., 2011) by identifying motivations of the majority parties and by prioritising them to understand the opening of an absolute majority. To do this, by identifying the motivations that resulted in the coalition of the 22 political groups with an absolute majority after the municipal elections of 14 October 2018, our research question comprises two elements: 1) we seek to analyse the different motivations that explain the negotiations of oversized coalitions whose main partner obtained an absolute majority, 2) and we do so in the municipal context in the Walloon Region.
  In relation to this second element, there is a factor that has not been sufficiently explored in the literature and that, in our opinion, is very important at the local level: this is the weight of interpersonal relations, which has long been identified (Gamson, 1961). Thus, several authors have emphasised the importance of social relations in the selection of coalition partners (Dumont et al., 2008, p. 133), particularly in view of their personality, trust, respect and mutual affinities (Wille & Deschouwer, 2012, p. 122). In this study, therefore, we seek a better understanding of the weight of interpersonal relations in the opening of an absolute majority.
  For this study, we chose a qualitative methodology, which we believe is the most appropriate for identifying the motivations behind opening an absolute majority.

• 2 The Study of All the Walloon Cases Concerned

  The 22 municipalities that are the focus of this study have different characteristics that depend on their geographical location, the number of inhabitants, the almost directly appointed mayor, the lists present and the size of the absolute majority (see Annex 1).
  We begin by detailing some of these characteristics. First, from a partisan point of view, 11 municipalities are led by a mayor affiliated to the Mouvement Réformateur (hereafter MR), seven are led by a mayor who is a member of the Parti Socialiste (hereafter PS) and four mayors are affiliated to the Centre Démocratique Humaniste (hereafter cdH)³x For the sake of clarity the parties may be positioned politically. Thus, the cdH is a Christian-democratic party, traditionally placed in the centre of the left-right axis. ECOLO is an environmentalist party, somewhat centre-left. The MR is a liberal party positioned on the right. The PS is a left-wing party.. Fourteen mayors are men and eight are women. Second, the majority parties are divided into two types of lists. Most of the lists studied are national lists (Dandoy, Dodeigne, Matagne & Reuchamps, 2013) whose candidates ran under the official name of the national party (MR, PS, cdH). However, some lists can be considered ‘quasi-national’ (Ibid.), insofar as they adopted a name different from their party’s, while clearly showing their reference to it. Examples of these lists are PSD@ (Andenne), NAP-MR (Rixensart) and MR-IC (Péruwelz). This category also includes lists whose members are affiliated to a party (Dodeigne, Close & Matagne, 2020). Third, the size of the absolute majority held by a party varies from one municipality to another. The majority is narrow when it is held by one or two councillors. It is wide when the majority party has at least three more councillors than half of the seats to be filled.⁴x Where there is an odd number of seats to be filled, half of them shall be rounded up. Finally, it should be noted that 20 majority groups formed a majority with only one additional partner, while the two remaining groups (the PS in Charleroi and the MR in Courcelles) included two additional partners.
  To identify the motivations of the political groups in the 22 municipalities studied to open their absolute majority, semi-structured interviews were conducted with the mayors or leaders of the parties holding the absolute majority. There are three reasons for this choice. First, semi-structured interviews are the most appropriate method of data collection for understanding the motivations behind opening an absolute majority and, above all, for understanding the selection of the coalition partner from the perspective of the analysis of interpersonal relations and oversized coalitions (Coman et al., 2016, p. 110; Olislagers, 2013, p. 171). Second, as the number of communes concerned by the opening of an absolute majority is limited, we decided to study all coalition negotiations. This qualitative study therefore aims to be exhaustive in terms of the cases it covers. Third, as the decision to open is mainly in the hands of the leader or the candidate with the highest number of personal votes, we gave priority to meeting these people. In total, 18 mayors and one president of the local party with the absolute majority were interviewed.⁵x Sixteen out of 18 interviews were conducted face-to-face, 1 interview was conducted by telephone, and 1 mayor wished to answer our questions in writing only. Two mayors never responded to our many requests, while one leader cancelled the appointment because of the Covid-19 health situation. To compensate for the lack of interviews and to enrich the data, local press articles published during the negotiations of the municipal majorities were collected.⁶x It should be noted that the negotiation period is indirectly regulated by the CDLD, as the draft agreement(s) is (are) deposited with the director general by the second Monday of November following the elections (i.e. 12 November 2018). Article L 1123-1 §2 CDLD. The articles in the local press dealing with the coalition negotiations between 15 October 2018 and 18 November 2018 were therefore collected.
  In addition to the interviews with the mayors and leaders of 19 political groups with an absolute majority, we also met five representatives of the support party who headed the list for the municipal elections and who became aldermen following the formation of the coalitions. We selected these five representatives on the basis of the communes in which the political parties had a majority, following both the communal elections of 14 October 2012 and the communal elections of 14 October 2018, but which only favoured opening following the second elections. The choice of this criterion stems from our desire to determine the potential triggers that led the representatives of these parties to open up their majority although they had not done so following the first election. Five municipalities meet this criterion: Aywaille, Bernissart, Grâce-Hollogne, Rixensart and Wavre.⁷x The interviews with these five aldermen were conducted by videoconference owing to the health situation related to Covid-19. We chose not to meet the representatives of the support party systematically for two reasons. To begin with, practical reasons prevented us from devoting resources and time to additional interviews, and, moreover, we were faced with a saturation of the data collected, in that additional interviews no longer seemed to allow, at a certain stage, the emergence of new information.
  In conducting the interviews, the questions focused exclusively on one stage of the ‘life cycle of political coalitions’ (Müller et al., 2008), namely the coalition building phase. All interviews were transcribed. As some representatives wished to remain anonymous, references to the interviews are anonymised in the presentation of the results.
  The method of analysis was qualitative content analysis, which is used to describe the meaning of the qualitative material studied (Schreier, 2012, pp. 1-19). To achieve this, we identified the paragraphs of the transcribed interviews referring to motivations related to the opening of the absolute majority. Our research question thus related to the angle from which we analysed our material. These paragraphs were coded into different categories according to their meaning. The coded paragraphs were gradually grouped into main themes, which are reflected in the structure of this article.⁸x Our coding thus fulfilled the requirements of a qualitative content analysis: unidimensionality of the codes (the coding should capture only one aspect of the material), mutual exclusivity of the codes, completeness of the codes and saturation of the codes (all codes should refer to a content of the material) (Schreier, 2012, pp. 71-79). Annex 2 presents the first codes used to categorise the paragraphs from the interviews, as well as the main themes that emerged from the comparative work and that served as the basis for the structure of this article. As it is we who decided on the intrinsic meaning of the analysed material, our analysis had to be systematic, from two points of view: we first examined all our material by identifying those parts of it that could be related to our research question and coding them consecutively according to their meaning. We then applied, the same sequence of steps to each interview in our material: reading, identifying paragraphs related to the research question, coding and grouping into main themes. This work was carried out with the help of a qualitative analysis software belonging to the computer-assisted qualitative data analysis (CAQDAS) family, namely Weft. This software promotes reflexivity (Woods, Macklin, Lewis, 2016) and, above all, enabled us to match the different motivations apparent in the interviews, on the basis of the codes we established. The coding was carried out in light of previous scientific findings. We regularly found traces of previous scientific studies in our qualitative content analysis. The identification of the different motivations explaining the opening of an absolute majority is thus the result of an iterative process (Sinkovics & Alfodi, 2012) between the results of previous research and our interviews.

• 3 Opening Local Absolute Majorities

  On the basis of our qualitative content analysis, we present the motivations behind opening an absolute majority at the local level. Specifically, our analysis identified two key stages in the formation of the coalitions studied; the first relates to the decision to open (3.1), while the second refers to the choice of the minority partner (3.2).

  3.1 The Choice to Open

  There are three main reasons for justifying the choice to open: to ensure more ‘efficient’ municipal management, to have additional relays within the other levels of power, and to form a majority that is more representative of the population, making it possible to give an image of openness, while confirming some of the results of previous research and at the same time enriching it.

  3.1.1 The Triple Efficiencies of Local Coalitions

  The choice to open can be explained, first of all, by the desire to ensure a more efficient local management, thus refining the results of previous studies on the fulfilment of ‘policy-seeking’ (Axelrod, 1970; Carrubba & Volden, 2000; Crombez, 1996; Dumont et al., 2008; Olislagers & Steyvers, 2015; Strom, 1990; Wille & Deschouwer, 2012). More specifically, our study allows us to categorise this efficiency because the interviews reveal three different registers of discourse that allow a political ideology to be fulfilled.
  First, efficiency is institutional. Seventeen of 22 political groups have a narrow absolute majority. This means that if one or two councillors from the majority group are absent from a meeting of the municipal council, the majority group loses its absolute majority and may not be able to adopt the items on the agenda. When a group within the municipal college has a number of councillors almost equivalent to the number of opposition members, the risk of blocking local management is considered too great by the actors interviewed (BA5, BD3, BX1, BA4, BA3, BX9, BX6, BA8 & PD9). The mayors indeed generally feel more comfortable carrying out the day-to-day work of a municipality when they hold a broader majority. This allows them to carry out the work more calmly and comfortably (BA5, BD3, BX1). Consequently, political groups with a narrow absolute majority prefer opening up to make it easier to run the municipality and to implement their electoral programme (PX2).
  The interviews allow us to observe that this desire to form a coalition to strengthen the majority may stem from two factors. First, the introduction of a motion of individual and collective distrust in the local law of the Walloon Region in December 2005 has, in some municipalities (BD3), accentuated this need to have a more comfortable majority. Since 2006, local majorities can effectively be overturned by “at least half of the councillors of each political group forming an alternative majority” (Article L1123-14 CDLD; Matagne et al., 2011; Gustin, 2018). Forming an oversized coalition thus reduces the risk of a vote of no confidence (BA3). Moreover, political formations want to form a comfortable majority because of past experience of political deadlock resulting from a narrow majority (BD3, BA4 & BX6).
  Second, efficiency is strategic. The choice to open is explained by the leaders’ desire to ensure a narrow minority in the local council (BX5, BD3, BX9 and BX4). Indeed, the consequence of a large majority is a smaller opposition. When the opposition is composed of a few people from one or two parties, it takes up little space, since criticism and remarks on the public policies adopted carry less weight. The aldermen and mayors say that they spend more time on their daily work and less on defending their projects before the population or the media (BX4). According to one mayor, a majority whose decisions are rarely contested has a better image among citizens (BX5).
  Third, efficiency is stabilising. If the search for stability has already been analysed at the Belgian local level (Nassaux, 2002), we can see that this stability feeds efficiency in two ways. To begin with, dissensions can exist between the members of the same political group, which is notably the case for the opening lists made up of several political tendencies. The formation of a coalition is then a means to reduce tensions within political groups (BA6, BA7, BX7 and BX9). In addition, when the absolute majority is narrower, the relative weight of each of its members is all the more significant. The choice to open may therefore result from the desire not to depend on a minority of one or two people exercising ‘pressure’, attempting to ‘blackmail’ or imposing ‘a veto’ (BA3, BA7, BD3, BX6, BX4, PX2).
  In the end, the three types of efficiency that justify opening an absolute majority contribute to the will to realise political ideologies through public policies, both by the party holding the absolute majority and by the coalition partner (BA3, BA5, BD2, PD6, PD9 and PD7), whether it is to have the necessary majority in the local council to adopt public policies, to ensure that the opposition is as narrow as possible and can criticise public policies or to guarantee a certain stability in the adoption of public policies.

  3.1.2 The Relays of Local Coalitions

  The choice to open is explained by the desire to achieve a certain congruence, as described previously. This congruence is particularly important for local political actors like the municipalities, acting as subordinate powers, who have discretionary decision-making power and organic autonomy under decentralisation but also exercise certain competences as representatives of federal, regional or community institutions under devolution (Articles 41 and 162 of the Belgian Constitution; Durviaux, 2018). In the context of implementing their competences, local political actors are therefore frequently in contact with higher levels of power, whether to ‘obtain information on the progress of files’, on decisions taken or on any other service (BA1, BD3, BA5, BX1, BD2, BA4, BA3, BX3, BA6, BX7, BX9 and BA2). In such an institutional configuration, members of the municipal colleges find it easier to communicate with political representatives at higher levels of their political colour (BA3, BX7 and BA7), thereby revealing the importance of interpersonal relations. Consequently, political groups with an absolute majority form a coalition with another political group present in the executives of higher levels of power (BX1, BA4, BD2, BX7, BX5, BA8, BX4, BA1, BA5 and PX2). It should be noted that, generally speaking, leaders appreciate being in contact with political representatives at these levels even when their party is not in the majority. These representatives are relays that help to implement the municipal electoral programme (BD3, BX3, BA6 and PC1).
  However, the desire to have relays in the higher levels of power is undermined by the election timetables. Indeed, municipal and provincial elections are held every six years in mid-October, while regional and federal elections are held every five years at the end of May (except in the case of early federal elections following a political crisis). Thus, when the local majorities were set up in 2018, the political groups did not know with certainty which parties would compose the federal, regional and community executives following the elections scheduled for 26 May 2019, although there were ‘rumours’ (BD8) about potential partners.
  In the end, opening up an absolute majority is justified by the importance of the relays available to local elected officials, allowing them to facilitate the realisation of the communal electoral programme, thus confirming the results of previous scientific studies on congruence in a multilevel political system (Bäck et al., 2013; Deschouwer, 2009; Strom, 1990), while still stressing the importance of interpersonal relations.

  3.1.3 The Representativeness of Local Coalitions

  The choice to open is ultimately justified by the desire to have a more representative majority of the population (AC3, BA1 and BX5). To better understand this argument, let us detail a distinctive feature of the local electoral system in Belgium. In the 22 municipalities that are the focus of this study, 17 political groups won a majority in seats, but not in votes. This can be explained by the voting system and, more precisely, by the Imperiali key used to transform votes into seats (Close & Matagne, 2020). This key ensures a more advantageous representation for the list with the best result (Lagasse, 2001). For example, the PS in Charleroi, with 41.29% of the votes, won 26 out of 51 seats, i.e. the majority of the seats to be filled in the municipal council. In view of the Imperiali key, some mayors wish to ensure a broader representation (BX9 and PX2).
  Opening an absolute majority is part of the desire of local political actors to project an image of openness (BA5). The leaders then mobilise a communication argument towards the citizens, demonstrating that they are listening to the electorate (BA3). In recent years, various municipalities have been affected by political and financial scandals that have tarnished the image of elected representatives (the Carolo, the Publifin affair and the Samusocial) (BD2). To improve this image, several mayors decided to stop governing alone and to share local power with other political groups (BX4, BA3, BA5, BD3, BA6, BA7 and PD8), thereby facilitating the gathering of a wider range of political information from citizens (BX5 and BX8) and providing food for thought for the members of the communal colleges (BD2, BX7, BX8, BA2 and BA8)
  Local political actors thus seem to have integrated the specificity of an institutional constraint (Strom et al., 1994) so as to further legitimise the decisions taken by the coalitions in place.

  3.2 The Choice of the Minority Partner

  Having presented the motivations that explain the opening up of an absolute majority, let us now detail the motivations that impel the leaders to choose one partner over another. Five were identified following the interviews with the leaders, allowing us to confirm some of the results of previous research while also enriching it.

  3.2.1 Ideological and Programmatic Proximity to the Minority Partner

  As various authors have already shown, by choosing to open an absolute majority, leaders ensure ideological and programmatic proximity with their majority partner (Adrian, 1977; Bué & Desage, 2009; Strom, 1990) and consequently form a ‘minimal connected winning coalition’ (Axelrod, 1970), thus making it easier to implement their electoral promises (Olislagers & Steyvers, 2015).
  While this reason is clearly present in the leaders’ speeches (BA5, BA7, BA2, BX1, BX3, BX6 and PX2), it is worth questioning the claim that political groups at local levels do not form ‘minimal connected winning coalitions’, preferring instead an ally that advocates similar local projects. To this end, we identified the associations favoured by the majority parties in the 22 municipalities studied (Table 1).

  Table 1 Numbers of Associations by Majority Parties

  		Majority Parties
  		PS	MR	cdH
  Minority Parties	PS		6	1
  	MR	3		3
  	cdH	2	3
  	Ecolo	3	3

  Note: It should be remembered that in two municipalities the majority party has joined forces with two partners, so the total is 24.

  It can be seen that the political groups do not systematically favour their ideological neighbour on a left-right axis. For example, the PS chooses the MR as many times as Ecolo (three times) and the MR is more likely to select the PS (six times) than the cdH (three times). There are two explanations for these associations. First, since certain majority political parties open their absolute majority to have relays at the higher levels of power, they tend to integrate a partner present at these levels of power (BX7 and BX5). Second, the proximity of the programmes at the local level should be highlighted. These programmes set out collaborative projects that are sometimes close between political groups, given the specificity of the competences exercised by the municipalities as subordinate powers. The different political groups (BA4, BA5, BX1, BX7 and BX6) do indeed share common ways of considering the local interest at the heart of these programmes, even if there are some differences in the way the programmes are implemented and even if differences in priorities and means of achieving projects may divide local parties (BX5, BA8, BA7, PX2 and PD9). However, let us not forget the historical rivalry between the two major Walloon political parties, the PS and the MR (BX7, BD3, BA1, BX3, PC1 and PD9)
  Although several interviews confirm the formation of minimal connected winning coalitions, it is clear that the specificity of local issues impels political actors to break away from them.

  3.2.2 Personal Affinities With the Minority Partner

  The choice of a coalition partner is based on personal affinities. This is certainly the main lesson that we could gather from our interviews (BA4, BA1, BD3, BX1, BA6, BX7, BX9, BA2 and BX4). According to several mayors, municipal politics, before being a relationship between political groups, is based on interpersonal relations (BA3, BA5, BX7, PC3 and PD9). From the interviews, we were able to identify five values that can strengthen these relations.
  First, trust is a key factor in the development of personal affinities. Members of a community college must trust and rely on each other (BA2, BA4, BX8, BA7 and PX2). Trust also means keeping commitments and keeping one’s word (BA1).
  Second, loyalty to the stronger party is an essential element for governing together, according to the leaders. The sense of loyalty is revealed, for example, by an opposition that was constructively led in the previous term (BX7 and BD7) or by the absence of aggression both during the election campaign and in the previous term (BA2 and PX2). The consequence of loyalty, according to some mayors, is a sense of collegiality (BA4, BA6 and PX2).
  Third, the way in which the day-to-day tasks are carried out must be in line with the way in which the future partner works; some prefer teamwork, while others have a more bilateral operation between the aldermen and the mayor (BX4).
  Fourth, the involvement of politicians and their willingness to collaborate are factors in the selection of the minority party (BX6, BA5, BA3 and BX7), especially based on past behaviours (BX1).
  Fifth, friendships may be formed within or beyond the local political sphere (BD5) and will have an impact on the choice of partner (BA4 and PX2).
  These five values, which contribute to the consolidation of interpersonal relationships, support the findings of previous research that has not sufficiently highlighted the values that underpin personal affinities in coalition building.

  3.2.3 Memories of Past Experiences With the Minority Partner

  Coalition building takes place in the ‘shadow of the past’ (Müller et al., 2008). While we have already identified the impact of negative past experiences in choosing to open up, positive past experiences have a greater impact on the choice of partner. Thus, given the work done in the past and the quality of interpersonal relations, the leaders of the majority party reaffirmed their willingness to work with the same partner again (BX7, BA2, BA8, BD3, BX1, BA4, BA6, BX8 and AC3). A few figures illustrate the importance of past positive experiences.
  Of the 22 municipalities, three have been governed by the same coalition for at least 18 years. In ten municipalities, the party that won an absolute majority renewed the alliance of the previous mandate, during which it had fewer elected members. In addition, in one municipality, the mayor chose his partner because he had the opportunity to work closely with someone who had been his representative when he was an alderman in the previous mandate (BA7).
  Finally, it should be noted that when two or even three parties formed a majority together during the previous mandate or when personal affinities linked politicians from different lists, pre-electoral negotiations took place, resulting either in simple ‘hallway’ chats (BX1, BA5, BA3, BA7 and PX2) or in an agreement that remained confidential (BA1, BD2, BA4, BX5 and BA6)

  3.2.4 The Number of Seats of the Minority Partner

  It has already been found that political parties tend to reduce the number of majority partners to obtain more mandates (Bonnet & Schemeil, 1972) and create ‘minimum winning coalitions’ (Riker, 1962). However, taking into account the number of seats held by the potential partner is not a prohibitive criterion (Olislagers & Steyvers, 2015). Our interviews confirm these results.
  On the one hand, when the leaders explicitly state that they take into account the number of seats obtained by the potential partners in 2018 (BD3, BA7, BX6, BA3, BX3, BX1, BX9 and PX2), it is because of a break in municipal politics: there has been a change either in mayors or in partners. In addition, these leaders had all won a narrow majority, prompting them to pay particular attention to the number of seats of the other political groups in two ways. First, some of them considered the fluctuation of seats and chose either the partner who won votes in the elections or the one whose results were stable. Second, the majority groups could also look at the number of seats obtained as such. From the interviews, we saw that political groups with a narrow absolute majority and a desire to extend this majority exclude lists with low representation, that is, those with only one or two elected members (BD3 a BA7), considering that one person is not enough to make the majority more stable (BA3). More generally, it should be noted that in 16 municipalities, the majority group formed a coalition with a group comprising at least three elected members, without necessarily choosing the groups with the most seats to avoid giving them too much visibility or having to give them several mandates in the municipal college (BX9, BA4, BA8, PD8).
  On the other hand, several mayors claimed that they did not take into account the number of seats obtained in 2018 by the potential partners. This is particularly the case for political groups bound by a pre-electoral agreement and not wishing to break it (BA4, BA1 and BD2), but it also applies to those with a majority considered sufficiently comfortable, without a partner, but who nevertheless decided to open to a party, regardless of its number of elected members (BA6 and BX5).

  3.2.5 The Influence of the Higher Party Authorities

  While previous research has emphasised the importance of coalitional congruence in multilevel systems (Bäck et al., 2013; Deschouwer, 2009; Strom, 1990), it is clear that the influence of the upper echelons of political parties (Dumont et al., 2008) has not been sufficiently considered at the local level. It should be noted that political parties are organised at several levels in French-speaking Belgium. There are three main levels of cdH, MR and PS: the national, the district and, finally, the local.⁹x There is also a provincial level within the MR. The higher authorities thus generally refer to the district president or the national president. On the basis of the interviews, we can distinguish three degrees of influence exercised by the higher authorities.
  First, there was a simple exchange of information between local elected representatives and representatives of higher authorities. For example, some mayors state that the higher authorities did not intervene in the designation of the partner (BX8, BX4, BA1 and BA3), apart from a transmission of information between the leaders who won an absolute majority and the presidents of the federations (BA1, BX5, BA8, BC2, BD2, BA3, BA7, BA2, PD7, PD9 and PD6). This exchange of information is intended to only provide political parties at the national level with an overview of local political majorities (BD2 and BX1). Local elected representatives therefore have a broad measure of autonomy (BA7).
  Second, some leaders explain that they have in the past experienced pressure from higher authorities in the form of ‘recommendations’ (BA4, BA6, BX6 and PD8). In such situations, the mayors claim that, regardless of these recommendations, local autonomy prevails (BA4, BA6 and BX6). In other words, other local considerations such as personal affinities or election results are taken into account (BX6).
  Third, an ‘imposition’, that is, an agreement made by the party federations at the higher levels, is formulated to the majority political group (BA8, BX2, BD3, BX3, BA7, BX1, PX2, PD5 and PD7). It should be noted, however, that it is difficult to assess how often such agreements are concluded, as the figures given during the interviews vary widely. Some speak of two or three agreements for the 262 municipalities in the Walloon Region (BX1), others of a much higher percentage (BA6).
  In the end, the three differentiated degrees of influence that we have identified through the interviews allow us to refine the understanding of the search for congruence in a multilevel system, like that of Belgium.

• 4 Towards a Model of Motivation

  By answering the two questions posed in this article, we are able to draw some more general lessons regarding the negotiation of oversized coalitions and the negotiation of such majorities at the local level in Wallonia.
  With respect to the negotiation of oversized coalitions, our research confirmed the results of previous research. Thus, by distinguishing the two stages of coalition formation (the choice to open and the choice of partner), we have shown that opening up absolute majorities to form oversized coalitions allows 1) policy-seeking, 2) accounting for political majorities at higher levels of power and seeking congruence, 3) integrating the institutional constraint of the voting system (Imperiali key), 4) moving away from minimal connected winning coalitions, 5) taking into account the shadow of the past and 6) avoiding dissension within political parties.
  As regards the negotiation of oversized coalitions at the local Walloon level, our study has revealed new facets. Thus, by distinguishing between the two stages of coalition formation, we can draw seven additional lessons. First, we categorised the efficiencies (institutional, strategic and stabilising) that enabled the political ideologies of local political actors to be realised. Second, we identified the five values that strengthened personal affinities at the local level (trust, loyalty, way of working, involvement and friendship), while finding that the shadow of the past is likely to strengthen the quality of interpersonal relations and foster formal or informal pre-election agreements. Third, we have identified the reasons why political partners move away from ‘minimal connected winning coalitions’, namely, the conclusion of pre-election agreements as well as the broad nature of the absolute majority. Fourth, in the Belgian multilevel system we identified the role of the higher party instances through three differentiated degrees of influence (information, recommendation and imposition).
  However, it is possible to go further in terms of insights by proposing a model to explain the motivations behind political negotiations of oversized coalitions at the local level when the main partner has an absolute majority of seats. To this end, we have attempted to prioritise the different motivations for opening an absolute majority and choosing the minority partner. Annex 3 contains a table showing, for each municipality, the presence or absence in the discourse of the actors interviewed of the three motivations relating to the choice to open and the five motivations relating to the choice of the minority partner. From this table, we can see that, as far as the choice to open is concerned, the search for greater representation is a necessary motivation to justify an oversized coalition. This motivation is indeed apparent in the speeches of the leaders of the 22 municipalities. As far as the choice of partner is concerned, personal affinities and memories of the past also seem to be necessary motivations to justify the choice of a minority partner. These motivations are indeed again apparent in the speeches of the leaders of the 22 communes.
  On the basis of our qualitative analysis and the prioritisation of these motivations, we propose to establish a model for negotiating oversized coalitions at the local level when the main partner has an absolute majority of seats (Figure 1). This model identifies two successive stages: first, there is the choice to open; second, there is the choice of the minority partner. For each stage, we identify ‘necessary motivations’ for negotiating oversized coalitions and ‘supporting motivations’, to use Olislagers’ compelling distinction (Olislagers, 2013, p. 198). As regards the former, if an oversized coalition is negotiated, then these motivations are necessary for such a negotiation. The latter, by contrast, support the former without being necessary for negotiating oversized coalitions at the local level when the main partner has an absolute majority of seats.

  The model of negotiating oversized majorities at the local level
  

• 5 Conclusion

  This study had a twofold objective: analysing the motivations that lead a political group to open its absolute majority at the local level and to choose a minority party. To this end, the 22 Walloon municipalities concerned by such an opening after the municipal elections of 14 October 2018 were studied through semi-structured interviews with the leaders of the different political groups in these municipalities. By categorising the different motivations underlying the choice to open and that of the minority partner, we established a model that identifies the necessary motivations, namely the search for greater representativeness, with regard to the choice to open and personal affinities and memories of the past with regard to the choice of partner. In addition, we identified the motivations that support the previous ones without conditioning the negotiation of an oversized coalition, namely the search for a certain efficiency and for relays in the higher levels of power with regard to the choice to open and the search for ideological proximity, the taking into account of the number of seats of the minority partner and of the influence of the higher instances with regard to the choice of the partner. By focusing on the process tracing of the negotiation of oversized majorities, our study makes a dual contribution to the scientific literature: first, it identifies the motivations (necessary and supporting) that explain the opening, and, second, it prioritises these motivations.
  On the basis of these results, it seems to us that four lines of research should be pursued. First, it would be useful to conduct comparative studies to identify potential specificities of the political contexts studied. For example, the importance of personal affinities could be explained by the lower nationalisation of local politics in Wallonia (Dodeigne et al., 2020). Some of the motivations identified in this article might be found in other local political contexts. Second, and further to the previous line of enquiry, it would be useful to develop a quantitative methodology to quantify, on larger scales, the importance of motivations for negotiating oversized coalitions, if a quantitative methodology is suitable, of which Ellen Olislagers is not convinced (2013, p. 171). This quantitative methodology would allow the results to be generalised by comparing the cases studied and distinguishing contextual factors. Third, it would now be necessary to identify the extent to which opening an absolute majority is advantageous for the majority group. Following the example of the study that identified the winners and losers of votes on a collective or individual motion of constructive mistrust over a period of two mandates (Gustin & Grandjean, 2019), it would be interesting to know the evolution of the results, over several mandates, of political groups that opted for opening their absolute majority. Fourth, it would be interesting to know the motivations of the partner parties to join oversized coalitions, again with the possibility of establishing a model of motivations. Such research would complement and enrich our research findings and identify some limitations.
  More broadly, this article demonstrates that the subjectivities of actors are an important consideration in the study of coalitions’ negotiations, thus inviting us to move away from a rational choice approach, on which many theories in this field of research are based.

References

• Adrian, C. (1977). Review of Coalition Theories: A Logical and Empirical Critique. The American Political Science Review, 71(4), 1612. doi: 10.1017/S0003055400269839.

• Andeweg, R. B. (2003). Political Psychology: Prospects and Potential. European Political Science, 2(2), 23-29. doi: 10.1057/eps.2003.4.

• Andeweg, R. B. (2011). From Puzzles to Prospects for Coalition Theory. In R. B. Andeweg, L. De Winter & P. Dumont (Eds.), Puzzles of Government Formation (pp. 190-203). Routledge.

• Axelrod, R. (1970). Conflict of Interest: A Theory of Divergent Goals with Applications to Politics (p. 216). Markham Pub. Co.

• Bäck, H. (2003). Explaining and Predicting Coalition Outcomes: Conclusions from Studying Data on Local Coalitions. European Journal of Political Research, 42(4), 441-472. doi: 10.1111/1475-6765.00092.

• Bäck, H. (2008). Intra-Party Politics and Coalition Formation. Party Politics, 14(1), 71-89. doi: 10.1177/1354068807081818.

• Bäck, H., Debus, M., Müller, J. & Bäck, H. (2013). Regional Government Formation in Varying Multilevel Contexts: A Comparison of Eight European Countries. Regional Studies, 47(3), 368-387. doi: 10.1080/00343404.2012.733072.

• Baron, D. P. & Diermeier, D. (2001). Elections, Governments, and Parliaments in Proportional Representation. Quarterly Journal of Economics, 116(3), 933-967. doi: 10.1162/00335530152466278.

• Bonnet, G. & Schemeil, Y. (1972). La théorie des coalitions selon William Rik. Essai d’application aux élections municipales françaises de 1965 et 1971. Revue française de science politique, 2, 269-282.

• Budge, I. & Laver, M. (1986). Office Seeking and Policy Pursuit in Coalition Theory. Legislative Studies Quarterly, 11, 485-506. doi: 10.2307/439930.

• Bué, N. & Desage, F. (2009). Le ‘monde réel’ des coalitions. Politix, 88(4), 7-37.

• Carrubba, C. J. & Volden, C. (2000). Coalitional Politics and Logrolling in Legislative Institutions. American Journal of Political Science, 44(2), 261-277. doi: 10.2307/2669309.

• Close, C. & Matagne, G. (2020). S’allier pour gagner, s’allier pour gouverner. Alliances préélectorales et coalitions locales en Wallonie et à Bruxelles. In J. Dodeigne, C. Close, V. Jacquet & G. Matagne (Eds.), Les élections locales du 14 octobre 2018 en Wallonie et à Bruxelles: une offre politique renouvelée? (pp. 125-148). Vanden Broele.

• Coman, R., Crespy, A., Louault, F., Morin, J-F., Pilet, J-B. & van Haute, É. (2016). Méthodes de la science politique. De la question de départ à l’analyse des données, Brussels, De Boeck, coll. “Méthodes en sciences humaines,” p. 224.

• Copus, C. (2019). Local Political Leadership. The Voters or Councilors - Who Chooses Who Governs? In R. Kerly, J. Liddle & P. T. Dunning (Eds.), The Routledge Handbook of International Local Government (pp. 25-37). Routledge.

• Crombez, C. (1996). Minority Governments, Minimal Winning Coalitions and Surplus Majorities in Parliamentary Systems. European Journal of Political Research, 29(1), 1-29. doi: 10.1111/j.1475-6765.1996.tb00639.x.

• Dandoy, R. (2018). Résultats des élections et formation des exécutifs. In F. Bouhon & M. Reuchamps (dir.), Les systèmes électoraux de la Belgique (2nd ed., pp. 513-534). Larcier.

• Dandoy, R., Dodeigne, J., Matagne, G. & Reuchamps, M. (2013). Les élections communales de 2012 en Wallonie (p. 195). Vanden Broele Editions.

• Debus, M. (2008a). Office and Policy Payoffs in Coalition Governments. Party Politics, 14(5), 515-538. doi: 10.1177/1354068807088121.

• Debus, M. (2008b). Party Competition and Government Formation in Multilevel Settings: Evidence from Germany. Government and Opposition, 43(4), 505-538. doi: 10.1111/j.1477-7053.2008.00267.x.

• Debus, M. & Gross, M. (2016). Coalition Formation at the Local Level: Institutional Constraints, Party Policy Conflict, and Office-Seeking Political Parties. Party Politics, 22(6), 835-846. doi: 10.1177/1354068815576292.

• Deschouwer, K. (2009). Coalition Formation and Congruence in a Multi-layered Setting: Belgium 1995-2008. Regional and Federal Studies, 19(1), 13-35. doi: 10.1080/13597560802692256.

• Diermeier, D. & Merlo, A. (2000). Government Turnover in Parliamentary Democracies. Journal of Economic Theory, 94(1), 46-79. doi: 10.1006/jeth.2000.2682.

• Dodd, L. (1974). Party Coalitions in Multiparty Parliaments: A Game-Theoretic Analysis. The American Political Science Review, 68(3), 1093-1117. doi: 10.2307/1959150.

• Dodeigne, J., Close, C. & Matagne, G. (2020). Le scrutin communal: une affaire nationale ou locale? In J. Dodeigne, C. Close, V. Jacquet & G. Matagne (Eds.), Les élections locales du 14 octobre 2018 en Wallonie et à Bruxelles: une offre politique renouvelée? (pp. 33-58).Vanden Broele.

• Downs, A. (1957). An Economic Theory of Political Action in a Democracy. The Journal of Political Economy, 65(2), 137-150. doi: 10.1086/257897.

• Downs, W. M. (1998). Coalition Government Subnational Style (p. 336). Ohio State University Press.

• Dumont, P., De Winter, L. & Ackaert, J. (2008). La formation des coalitions communales: les pratiques anciennes en héritage. In B. Rihoux, J. Buelens & K. Deschouwer (Eds.), Entre l’électeur et le quartier général. Les sections locales des partis et les élections communales de 2006 (p. 206). VUB Press.

• Dumont, P., De Winter, L. & Andeweg, R. B. (2011). From Coalition Theory to Coalition Puzzles. In R. B. Andeweg, L. De Winter & P. Dumont (Eds.), Puzzles of Government Formation (pp. 1-23). Routledge.

• Durviaux, A. L. (2018). Principes de droit administratif (p. 672). Larcier.

• Ervik, B. (2015). Does the Mayor’s Party Brand Matter? The Prospect of Statewide Parties’ Policy in Local Government. Local Government Studies, 41(3), 339-361. doi: 10.1080/03003930.2013.836493.

• Gamson, W. (1961). A Theory of Coalition Formation. American Sociological Review, 26(3), 373-382. doi: 10.2307/2090664.

• Grandjean, G. (2016). Onderhandelen over vorming van een meerderheid en burgemeester worden in België: Welke invloed heeft de regionale wetgeving? Publiekrechtelijke Kronieken, 4, 594-610.

• Gustin, A. (2018). Les motions de méfiance constructives dans les communes wallonnes (2012-2018). Courrier hebdomadaire du CRISP, 2378, 5-36.

• Gustin, A. & Grandjean, G. (2019). Politieke gevolgen van het indienen van constructieve moties van wantrouwen op gemeentelijk niveau (2006-2018). Publiekrechtelijke Kronieken, 1, 45-61.

• Jungar, A-C. (2011). The Rainbow Coalition. A Surplus Majority Coalition in Finland. In R. B. Andeweg, L. De Winter & P. Dumont (Eds.), Puzzles of Government Formation (pp. 129-146). Routledge.

• Klüver, H. & Bäck, H. (2019). Coalition Agreements, Issue Attention, and Cabinet Governance. Comparative Political Studies, 52(13-14), 1995-2031. doi: 10.1177/0010414019830726.

• Lagasse, N. (2001). Les règles particulières aux élections communales et les spécificités locales. In Centre de droit public, Les élections dans tous leurs états. Bilan, enjeux et perspectives du droit électoral (pp. 293-342). Bruylant.

• Laver, M. (1989). Theories of Coalition Formation and Local Government Coalitions. In C. Mellors & B. Pijnenburg (Eds.), Political Parties and Coalitions in European Local Government (pp. 15-33). Routledge.

• Laver, M., Rallings, C. & Thrasher, M. (1987). Coalition Theory and Local Government: Coalition Payoffs in Britain. British Journal of Political Science, 17(4), 501-509. doi: 10.1017/S0007123400004890.

• Laver, M. & Schofield, N. (1990). Multiparty Government: The Politics of Coalition in Europe (p. 336). Oxford University Press.

• Lijphart, A. (1984). Democracies: Patterns of Majoritarian and Consensus Government in Twenty-One Countries (p. 232). Yale University Press.

• Lupia, A. & Strom, K. (2008). Bargaining, Transaction Costs, and Coalition Governance. In K. Strom, W. C. Müller & T. Bergman (Eds.), Cabinets and Coalition Bargaining (pp. 51-53). Oxford University Press.

• Magre, J. & Pano, E. (2019). The Architecture of the Local Political Community. France, Italy, Portugal and Spain. In R. Kerly, J. Liddle & P. T. Dunning (Eds.), The Routledge Handbook of International Local Government (pp. 418-431). Routledge.

• Matagne, G., Radoux, E. & Verjans, P. (2011). La composition du collège communal après la réforme du Code wallon de la démocratie locale. Courrier hebdomadaire du CRISP, 2094, 5-35.

• Müller, W. C., Bergman, T. & Strom, K. (2008). Coalition Theory and Cabinet Governance: An Introduction. In K. Strom, W. C. Müller & T. Bergman (Eds.), Cabinets and Coalition Bargaining: The Democratic Life Cycle in Western Europe (pp. 1-50). Oxford University Press.

• Müller, W. C. & Strom, K. (2008). Coalition Agreements and Cabinet Governance. In K. Strom, W. C. Müller & T. Bergman (Eds.), Cabinets and Coalition Bargaining: The Democratic Life Cycle in Western Europe (pp. 159-199). Oxford University Press.

• Nassaux, J-P. (2002). La formation des coalitions dans les communes bruxelloises. Courrier hebdomadaire du CRISP, 1770, 5-39.

• Olislagers, E. (2013). Calculation or Donation? Choosing Partners in Local Coalition Formation in the Flemish Region of Belgium, Doctoral dissertation, Ghent University, p. 272.

• Olislagers, E. & Steyvers, K. (2015). Choosing Coalition Partners in Belgian Local Government. Local Government Studies, 41(2), 202-219. doi: 10.1080/03003930.2014.884496.

• Riker, W. (1962). Theory of Political Coalitions (p. 312). Yale University Press.

• Schreier, M. (2012). Qualitative Content Analysis in Practice (p. 272). Sage.

• Serritzlew, S., Skjæveland, A. & Blom-Hansen, J. (2008). Explaining Oversized Coalitions: Empirical Evidence from Local Governments. The Journal of Legislative Studies, 14(4), 421-450.

• Sinkovics, R. R. & Alfodi, E. A. (2012). Progressive Focusing and Trustworthiness. The Enabling Role of Computer-Assisted Qualitative Data Analysis Software (CAQDAS). Management International Review, 52, 817-845. doi: 10.2307/41682288.

• Sjölin, M. (1993). Coalition Politics and Parliamentary Power (p. 230). Lund University Press.

• Steyvers, K., Reynaert, H., De Ceuninck, K. & Valcke, T. (2008). All Politics Is Local, Partisan or National? Local Lists in Belgium. In M. Reiser & E. Holtmann (Eds.), Farewell to the Party Model? (pp. 169-194). VS Verlag für Sozialwissenschaften.

• Strom, K. (1990). A Behavioral Theory of Competitive Political Parties. American Journal of Political Science, 34(2), 565-598. doi: 10.2307/2111461.

• Strom, K., Budge, I. & Laver, M. J. (1994). Constraints on Cabinet Formation in Parliamentary Democracies. American Journal of Political Science, 38(2), 303-335. doi: 10.2307/2111406.

• Strom, K. & Swindle, S. M. (2002). Strategic Parliamentary Dissolution. American Political Science Review, 96(3), 575-591. doi: 10.1017/S0003055402000345.

• Thrasher, M. (1999). Sub-national Coalition Government Formation in Belgium, France and Germany. Local Government Studies, 25(1), 99-104. doi: 10.1080/03003939908433940.

• Volden, C. & Carrubba, C. J. (2004). The Formation of Oversized Coalitions in Parliamentary Democracies. American Journal of Political Science, 48(3), 521-537. doi: 10.2307/1519914.

• Wille, F. & Deschouwer, K. (2012). À propos d’hommes et de pouvoir: La formation des coalitions au sein des communes belges (p. 192). Academic and Scientific Publishers.

• Woods, M., Macklin, R. & Lewis, G. K. (2016). Researcher Reflexivity: Exploring the Impacts of CAQDAS Use. Social Research Methodology, 19(4), 385-403. doi: 10.1080/13645579.2015.1023964.

Annex 1 Characteristics of the Municipalities Studied
Table A1

Municipality	Province	Population (1 January 2018)	Majority Party	Partner Party	Percentage of the Majority Party Vote (%)	Number of Seats of the Majority Party	Number of Seats of the Partner Party	Short Majority (0) Large Majority (1)
Andenne	Namur	26,985	PSD@	MR	52.52	17/29	3	1
Ans	Liege	28,237	PS	MR-IC	45.32	16/29	6	0
Aywaille	Liege	12,393	Together (MR)	Progress (PS)	42.52	12/23	2	0
Bernissart	Hainaut	11,868	PS	Ecolo	46.66	11/21	2	0
Binche	Hainaut	33,590	PS	MR-IC (MR-cdH)	57.00	20/31	5	1
Charleroi	Hainaut	201,327	PS	C+ (cdH); Ecolo	41.29	26/51	4 + 3	0
Courcelles	Hainaut	31,309	MR	cdH; Ecolo	50.55	18/31	1 + 3	0
Estinnes	Hainaut	7,716	EMC (cdH)	MR	46.28	10/19	2	0
Florennes	Namur	11,365	Contact 21 (cdH)	Ad 11 (MR)	45.02	11/21	7	0
Grâce-Hollogne	Liege	22,524	PS	Ecolo	42.11	15/27	3	0
Ham-sur-Heure-Nalinnes	Hainaut	13,532	MR	VivrEnsemble (cdH)	69.73	18/23	2	1
Marche-en-Famenne	Luxembourg	17,455	cdH	PS	47.48	14/25	4	0
Montigny-le-Tilleul	Hainaut	10,136	MR	PS	47.35	12/21	3	0
Mouscron	Hainaut	58,164	cdH	MR	47.45	19/37	5	0
Péruwelz	Hainaut	17,103	MR-IC	Ecolo	43.30	13/25	2	0
Rixensart	Walloon Brabant	22,381	NAP-MR (MR-cdH)	SolidaRix (PS)	44.82	14/27	3	0
Seneffe	Hainaut	11,267	LB (MR-Defis-independent)	Ecolo	47.44	11/21	3	0
Silly	Hainaut	8,403	LB (MR)	SENS (cdH)	53.81	12/19	4	1
Spa	Liege	10,371	MR	SPA (PS)	43.79	11/21	1	0
Theux	Liege	12,027	IFR (MR)	PS Plus	45.10	12/23	4	0
Thuin	Hainaut	14,683	PS	IC (cdH)	41.88	12/23	4	0
Wavre	Walloon Brabant	34,310	LB (MR)	PS	40.63	16/31	3	0

Annex 2 The List of Codes and Themes
Table A2

Codes for Categorising Paragraphs
Aldermen	Historic alliance	Personal affinities
Autonomy	Internal legitimacy	Personality of the leader
Campaign	Internal party life	Positive experiences
Carolo	Large majority	Pre-election discussions
Challenges	Local context	Pressures
Choice of partner	Long term	Proximity of programmes
Comfortable majority	Maintaining the coalition	Regional or federal context
Communal tensions	Negative experiences	Relay
Competencies	Number of partners	Representation
Culture	Number of seats	Stable majority
Current term of office	Opening image	Tradition of opening
Electorate	Opposition	Trust
Enriching the debate	Partisan independence	Type
General political context	Partner	Workers
Higher bodies
Main Themes
Large majority	Internal party life	Past experiences
Choice of partner	Mandate	Personal relationships
Context	Municipal mandates and derivatives	Pre-election discussions
Distribution of competences	Negotiations	Relations with higher levels of power
Electorate	Opposition	Representation
Enriching the debate	Partner’s vision

Annex 3 The Presence of the Motives for Opening an Absolute Majority and Selecting the Minority Partner
Table A3

Municipalities	The Choice of Opening			Total	The Choice of the Minority Partner					Total
	Efficiencies	Relays	Representativeness	/3	Ideological Proximity	Personal Affinities	Memories of the Past	Number of Seats	Influence of Higher Authorities	/5
Andenne	1	1	1	3	1	1	1	1	0	4
Ans	1	1	1	3	1	1	1	1	0	4
Aywaille	1	1	1	3	1	1	1	1	0	4
Bernissart	1	0	1	2	1	1	1	1	1	5
Binche	0	0	1	1	0	1	1	1	0	3
Charleroi	1	1	1	3	0	1	1	1	0	3
Courcelles	1	1	1	3	0	1	1	1	1	4
Estinnes	1	1	1	3	0	1	1	1	1	4
Florennes	1	0	1	2	0	1	1	1	1	4
Grâce-Hollogne	0	0	1	1	1	1	1	0	1	4
Ham-sur-Heure-Nalinnes	1	1	1	3	1	1	1	0	0	3
Marche-en-Famenne	0	0	1	1	0	1	1	0	0	2
Montigny-le-Tilleul	1	1	1	3	1	1	1	0	0	3
Mouscron	1	1	1	3	0	1	1	1	1	4
Péruwelz	1	1	1	3	0	1	1	1	1	4
Rixensart	1	0	1	2	0	1	1	1	1	4
Seneffe	1	0	1	2	0	1	1	1	0	3
Silly	1	1	1	3	1	1	1	0	1	4
Spa	0	1	1	2	1	1	1	0	1	4
Theux	1	1	1	3	0	1	1	0	1	3
Thuin	1	1	1	3	1	1	1	1	0	4
Wavre	1	1	1	2	1	1	1	0	1	3
Total	18	15	22		11	22	22	14	12

Noten

• * This article is published with the support of Cité (ULiège).

• 1 The majority pact includes the indication of the political groups that form the municipal majority, the identity of the mayor, the aldermen and the president of the social action council. Article 1123-1 et seq. of the Code of Local Democracy and Decentralisation (hereafter CDLD).

• 2 The 22 communes are: Andenne, Ans, Aywaille, Bernissart, Binche, Charleroi, Courcelles, Estinnes, Florennes, Grâce-Hollogne, Ham-sur-Heure-Nalinnes, Marche-en-Famenne, Montigny-le-Tilleul, Mouscron, Péruwelz, Rixensart, Seneffe, Silly, Spa, Theux, Thuin and Wavre.

• 3 For the sake of clarity the parties may be positioned politically. Thus, the cdH is a Christian-democratic party, traditionally placed in the centre of the left-right axis. ECOLO is an environmentalist party, somewhat centre-left. The MR is a liberal party positioned on the right. The PS is a left-wing party.

• 4 Where there is an odd number of seats to be filled, half of them shall be rounded up.

• 5 Sixteen out of 18 interviews were conducted face-to-face, 1 interview was conducted by telephone, and 1 mayor wished to answer our questions in writing only.

• 6 It should be noted that the negotiation period is indirectly regulated by the CDLD, as the draft agreement(s) is (are) deposited with the director general by the second Monday of November following the elections (i.e. 12 November 2018). Article L 1123-1 §2 CDLD. The articles in the local press dealing with the coalition negotiations between 15 October 2018 and 18 November 2018 were therefore collected.

• 7 The interviews with these five aldermen were conducted by videoconference owing to the health situation related to Covid-19.

• 8 Our coding thus fulfilled the requirements of a qualitative content analysis: unidimensionality of the codes (the coding should capture only one aspect of the material), mutual exclusivity of the codes, completeness of the codes and saturation of the codes (all codes should refer to a content of the material) (Schreier, 2012, pp. 71-79).

• 9 There is also a provincial level within the MR.