Investigating the gendered and institutional dimensions of the opportunity structures within political parties is essential in order to explain women’s chronic minority status in politics as well as the persistence of male dominance. The small but growing body of work on gender, political parties and candidate selection has contributed many important insights about the dynamics of these processes and how they should be studied (see, for example, Bjarnegård Reference Bjarnegård2013; Hinojosa Reference Hinojosa2012; Kenny Reference Kenny2013; Kittilson Reference Kittilson2006; Lovenduski and Norris Reference Lovenduski and Norris1993; Murray Reference Murray2010; Norris and Lovenduski Reference Norris and Lovenduski1995). That political parties are gatekeepers and should therefore be central to the analyses is one important starting-point. Another significant insight is that gender does not only operate at the individual level – we can also understand and analyse gender at the institutional level and therefore critically examine political parties as gendered organizations. A third insight has to do with the importance of informal rules in the institutional setting that parties constitute or, rather, the interaction between formal regulations and informal practices at the party level. Research has demonstrated, usually through detailed case studies, how all these insights are important in order to capture what really matters when parties select candidates. Yet, while the important role of political parties in shaping patterns of women’s representation is widely recognized, there have been surprisingly few systematic comparative studies into the ‘secret garden’ of candidate selection and recruitment.
The aim of this article is to move a step forward by assessing the comparative potential in the field, focusing on parties as gendered organizations that are guided by both formal and informal rules. We propose an ambitious but realistic research agenda for how to analyse the relationship between gender, institutions and candidate selection in comparative perspective. We begin by outlining in greater detail the theoretical and methodological challenges of researching the gendered and comparative dynamics of the candidate selection process. We then move on to explore the possibilities of comparing these complex dynamics by revisiting original research on the candidate selection process in two different settings – Thailand and Scotland – and two different political parties – the Thai Rak Thai and the Scottish Labour Party. We situate these earlier studies in the context of our comparative research agenda, highlighting two key aspects of the candidate selection process that have gendered consequences in both cases: the role of party recruitment and the role of local influence over selection decisions. We elaborate on how and why these mechanisms are gendered and demonstrate their importance for understanding the continuity of male over-representation and female under-representation in both contexts. Building on these commonalities, and elaborating on what needs to be studied, and how, the article concludes by setting out a research agenda for future work in the field.
CHALLENGES IN THE COMPARATIVE STUDY OF GENDER AND CANDIDATE SELECTION
Research seeking to understand male over-representation in politics has increasingly started focusing on political parties as gendered organizations, investigating the specific set of institutional conditions under which women can achieve concrete gains (Bjarnegård Reference Bjarnegård2013; Kittilson Reference Kittilson2013). In this view, gender operates both as a (socially constructed) category and as a feature of institutions and social structures (Krook and Mackay Reference Krook and Mackay2011). So, for example, studies of gender and candidate selection demonstrate that gender plays out at the individual level through direct or indirect discrimination by party gatekeepers (Norris and Lovenduski Reference Norris and Lovenduski1995). But they also highlight that these gendered interactions take place at the party level within a framework of both formal and informal party rules and practices that are shaped and structured by gender norms – favouring the model of the ‘ideal candidate’, who is usually a man (Chapman Reference Chapman1993; Lovenduski and Norris Reference Lovenduski and Norris1989).
Establishing a clear picture of internal party dynamics, therefore, requires a gendered lens. Parties are gendered organizations, in that they are characterized by traditional (and often unacknowledged) conceptions of gender relations that generally disadvantage women (Lovenduski Reference Lovenduski2005). There is, consequently, an increasing amount of empirical contributions specifically focusing on gendered aspects of candidate selection and recruitment (Bjarnegård Reference Bjarnegård2013; Freidenvall Reference Freidenvall2006; Hinojosa Reference Hinojosa2012; Kenny Reference Kenny2013; Kittilson Reference Kittilson2006; Lovenduski and Norris Reference Lovenduski and Norris1993; Murray Reference Murray2010; Norris and Lovenduski Reference Norris and Lovenduski1995). A key first step in this research agenda is to map and analyse the ‘gender regimes’ of political parties, starting with the formal architecture and informal rules, norms and practices of the selection process (cf. Lovenduski Reference Lovenduski2011). It is, however, often difficult to access information about internal selection processes, partly because of the often informal and ‘hidden’ character of these practices. Much of the research in this area has, therefore, continued to focus on formal regulations and official party rules, often at the expense of exploring the informal aspects of candidate selection and recruitment and their gendered consequences.
This is not to argue that formal rules do not matter for gender and candidate selection; rather, formal rules should be understood in connection to the informal practices that they affect and are affected by (cf. Bjarnegård and Kenny Reference Bjarnegård and Kenny2015). There is a large literature demonstrating how formal regulations such as electoral systems and electoral gender quotas can fundamentally shape and alter party selection practices in gendered ways. Rules such as electoral systems matter because they provide political parties with incentives that have an impact on who parties perceive to be a suitable candidate. Candidate gender quotas, on the other hand, matter because parties are required to put gender on the table and design their selection processes in such a way that they are able to identify not only suitable candidates in general (who often turn out to be male) but also suitable female candidates. The impact of other formal rules is less clear. For instance, there has been an inconclusive discussion about whether centralized or decentralized selection procedures favour women. While some argue that a decentralized party structure may lead to gains for women at the grassroots level (Lovenduski and Norris Reference Lovenduski and Norris1993; Norris Reference Norris1996, Reference Norris1997), others suggest that centralized party organizations give elites more power to implement and enforce gender equality reforms – when they are willing to do so (Kenny and Verge Reference Kenny and Verge2013; Kittilson Reference Kittilson2006; Murray Reference Murray2010).
The reason for these inconclusive accounts is probably that the above-mentioned studies do not necessarily investigate the same things. For example, in seeking to answer the question of ‘who decides’ in the candidate selection process, and the consequences of this for women’s representation, studies have largely focused on where decision-making takes place, whether nationally or locally (Kenny and Verge Reference Kenny and Verge2013; Kittilson Reference Kittilson2006), and on the degree of inclusiveness of the selectorate (Hinojosa Reference Hinojosa2012; Rahat and Hazan Reference Rahat and Hazan2001). Yet often these different dimensions are lumped together into broader discussions of ‘centralization’ that focus almost exclusively on formal rules and regulations, or look only at particular stages of the process (or all of them at once) (cf. Bjarnegård Reference Bjarnegård2013). Candidate selection is not static, however; it is a complex and temporally specific process that takes place in many steps, and formal rules on where decisions about candidates are taken do not always correspond to informal practices and de facto decisions taken at different levels. Reading and analysing party regulations is not enough even to understand whether and to what extent formal rules guide the selection process. We need instead to study the process that is shaped by the actual practices taking place within a specific formal framework (Bjarnegård and Kenny Reference Bjarnegård and Kenny2015). Where selection procedures are bureaucratized and in practice guided by a strong regulatory framework, the process for selecting candidates is not only described in some detail in party documents, but what is de jure described in the party regulations is also implemented – that is, de facto how the process for selecting candidates is carried out (Bjarnegård and Zetterberg Reference Bjarnegård, Kenny and Zetterberg2016; Norris Reference Norris1996). Understanding formal rules therefore necessitates comparing their content to actual practices. Determining how strong formal rules are, and to what extent they actually guide how candidate selection is done ‘on the ground’, is therefore one of the first steps towards understanding what leeway the formal framework leaves for informal practices to play a part in candidate selection (Bjarnegård Reference Bjarnegård2013; Bjarnegård and Kenny Reference Bjarnegård and Kenny2015; Bjarnegård and Zetterberg Reference Bjarnegård, Kenny and Zetterberg2016; Kenny Reference Kenny2013).
However, the possibilities of specifying and generalizing the gendered impact of different types of selection procedures across parties and countries are still limited due to the relative scarcity of systematic comparisons that take both formal and informal rules into account. Comparative studies that do exist generally take the form of anthologies where individual contributions on candidate selection in different countries are more or less explicitly related to the common theme of the book, but where there are few systematic and integrated comparisons (see, for example, Gallagher and Marsh Reference Gallagher and Marsh1988; Lovenduski and Norris Reference Lovenduski and Norris1993; Norris Reference Norris1997; Siavelis and Morgenstern Reference Siavelis and Morgenstern2008). Studies have, however, usefully compared candidate selection structures and gendered consequences of different political parties operating within the same country (see, for example, Bjarnegård Reference Bjarnegård2013; Freidenvall Reference Freidenvall2006; Kenny Reference Kenny2013; Murray Reference Murray2010; Norris and Lovenduski Reference Norris and Lovenduski1995; Verge and de la Fuente Reference Verge and de la Fuente2014). Such comparisons are well posed to chisel out behavioural differences between parties operating under similar circumstances. However, gendered analyses of candidate selection have, for the most part, stopped short of comparisons across parties and countries. Notable exceptions include Caul Kittilson (Reference Kittilson2006), who studied political parties and women’s representation in Western Europe, and Hinojosa (Reference Hinojosa2012), who compared Latin American political parties and their candidate selection procedures. Very few, if any, comparative cross-country studies of gendered aspects of candidate selection have included informal aspects of the selection process. Hinojosa (Reference Hinojosa2012: 12–13) explicitly addresses the informal nature of candidate selection in Latin America, though she acknowledges the difficulties of obtaining information about informal party practices, given the constant rule changes that take place in the region and the difficulty of obtaining reliable data from parties.
Methods for collecting comparative data on candidate selection, then, are not always straightforward, particularly as we move away from focusing only on formal procedures and collecting written material, such as party regulations (although these are not always readily available either!) to focusing on informal practices and conventions that are not written down. Gallagher and Marsh’s (Reference Gallagher and Marsh1988) characterization of candidate selection as the ‘secret garden of politics’ therefore still seems relevant and important, particularly for a gendered analysis. Candidate selection is a crucial activity in political parties, but because of the internal power struggles it entails, it is also often considered internal business. When studying internal party politics one deals with events that are normally not of a public character (Nelson Reference Nelson2005: 2) and ‘political practices that some people would prefer remain undocumented’ (Arghiros Reference Arghiros2001: viii). Moreover, while informal criteria are important for who becomes a politician, really understanding how informality matters in a certain country requires contextual knowledge and expertise in a way that makes it difficult to quickly access this information for a large number of countries. Recent work on gender and informal institutions more broadly has attempted to overcome these methodological challenges by drawing on methods from other areas of the social sciences, including institutional ethnography (Chappell and Waylen Reference Chappell and Waylen2013). When the aim is comparative, however, these sorts of methods are not always feasible, and ethnographic data collection does not easily lend itself to structured comparison. There are also particular issues that arise in dealing with elite political organizations such as parties, which may be reluctant to grant access to particular research settings or information, and where access may change over time (for example, if a party is in opposition or in government) (Kenny Reference Kenny2014; see also Lovenduski Reference Lovenduski2016, in this issue).
COMPARING THE ‘SECRET GARDEN’: CANDIDATE SELECTION IN THAILAND AND SCOTLAND
This article constitutes a first exploration of the possibilities and limits of cross-country comparison by comparing gendered aspects of candidate selection in two contexts that are seemingly very different: Scotland and Thailand. The empirical data used for this comparison was not gathered in the synchronized manner that a truly comparative research design would require. Instead, we bring together two existing case studies and revisit data gathered separately, but with a common analytical frameworkFootnote 1 focusing on the gendered and institutional dynamics of the candidate selection process. The aim of this exercise is to investigate what common insights can be pulled from these two cases, to identify fertile ground for future research, and also highlight limitations and/or challenges for comparative research in this field. We have been careful to make sure that the similarities we find draw on comparable stages of the election process. The fact that certain commonalities stand out, despite the different contexts and research strategies, can also be seen as an advantage. While there are limits to what we can generalize from this comparison, it can help us identify common causal mechanisms (of power, of continuity, of change), which can then be explored in future research in other contexts (cf. Mackay et al. Reference Mackay, Kenny and Chappell2010; Pierson Reference Pierson2004).
In particular, in order to capture what matters, we have explicitly focused on both formal and informal aspects of the candidate selection process, and the interplay between them (cf. Bjarnegård and Kenny Reference Bjarnegård and Kenny2015). We have done this by focusing our research on how recruitment took place, and, with the narratives of actual candidates and other party actors at hand, we have been able to revisit theories about party demand as well as the impact of party decentralization. Both cases were investigated with a time-consuming and field-intensive process-tracing method including semi-structured interviews with candidates and party gatekeepers. To unearth the ‘real’ rules that shape the selection process (both formal and informal), we talked to actors themselves about ‘how things are done around here’ and ‘why do you do X but not Y?’ (Lowndes Reference Lowndes2005: 306; see also Kenny Reference Kenny2014). Our interviews focused on the personal experiences of party gatekeepers and candidates and thus concerned issues that our respondents were well placed to answer. We asked candidates to tell us how they ended up as candidates, what they had needed to do to get there, what their major assets were, and what hurdles they had encountered. Party gatekeepers were asked to tell us how they reasoned about a particular candidate, why they ended up with one candidate instead of another, to what extent the decision was in their hands, and what influenced their decision. The narratives emanating from these kinds of interviews give a surprisingly clear picture of what is at stake. And while these kinds of methods may also produce divergent accounts from different participants, these interpretations are in themselves a part of the ‘process’, shedding light on the ways in which particular events and meanings were constructed at different times and in different institutional sites (see Kenny Reference Kenny2013 on this point). Thus, starting with these process descriptions, rather than with specifying formal institutional differences in how candidates are selected, is a fruitful way forward for comparative research, allowing us to ‘see’ the ways in which the rules of the selection process (both formal and informal) play out on the ground.
In both cases, a multi-stage analysis of the selection process was also conducted, tracing the full potential chain from being eligible to becoming an aspirant to being selected as the official candidate of the party to being elected as an MP.Footnote 2 Most work in the field, in contrast, takes only a partial look at the selection process, which, as we highlighted above, can sometimes lead to inconclusive results. If our goal is to identify gendered processes, though, we need to look at the entire chain (from application to election) in order to identify the precise stages at which under-representation begins to occur (Ashe and Stewart Reference Ashe and Stewart2012; Norris and Lovenduski Reference Norris and Lovenduski1995). Each step involves different actors, points to different characteristics and may require different methodological tools, but they also influence one another. Although a dynamic multi-stage approach makes for potentially complex comparisons, it is nevertheless a challenge worth accepting.
In developing a comparative research agenda on gender and candidate selection, we also need to consider which types of cases we compare, and what the comparison will contribute. In this analysis, we have opted to revisit the findings of two very different cases, in order to search for similiarities that can be theoretically and empirically explored in future research. The two parties we are looking at are the Thai Rak ThaiFootnote 3 in Thailand and the Labour Party in Scotland. Our analysis of the candidate selection process focuses on the Thai Rak Thai candidate selection procedures for constituency seats preceding the 2005 parliamentary elections in Thailand and the Scottish Labour Party candidate selection procedures for constituency seats in the run-up to the 2007 Scottish Parliament election.Footnote 4 Both these parties can be situated within a larger framework of political reform, albeit in different ways. Both were either newcomers to the political arena, or operating in a new political arena, with change high on the agenda.
The Scottish Labour Party is a social democratic political party in Scotland which operates as the regional section of the UK-wide Labour Party. Processes of institutional and constitutional restructuring in the 1990s created a ‘new’ institutional context for the party, with the creation of new parliamentary spaces and structures of governance in Scotland and Wales, as part of the partial devolution of power from the UK Parliament. The establishment of the Scottish Parliament in 1999 opened up unprecedented opportunities for innovation in the candidate selection, creating pressures that were acutely felt in the Scottish Labour Party, which had historically been markedly less hospitable to women candidates and officeholders than the party at the British level (Brown Reference Brown2001; Mackay Reference Mackay2004). Inside the party, women took advantage of the opportunities presented by devolution and were important players in arguments over the use of gender quota measures for Scottish Parliament elections. These internal party debates were heavily influenced by wider agendas, most notably party modernization and centralization of candidate selection procedures (Bradbury et al. Reference Bradbury, Mitchell, Bennie and Denver2000). As a result, the use of quotas was also supported by key men in the party, and Scottish Labour was ultimately the only party in Scotland to implement formal quota measures prior to the first elections to the Scottish Parliament in 1999, using a mechanism called ‘twinning’ in constituency seats and placing women in favourable positions on the party’s regional lists, which resulted in a parliamentary group that was 50 per cent women. In subsequent elections, the party continued to lead on women’s representation in Scotland and achieved equal representation or better in its parliamentary group until 2011. Underlying trends, however, suggest that this outcome was largely due to the power of incumbency after 1999, rather than the institutionalization of gender balance, and the underlying pattern of turnover suggests a re-masculinization of Labour candidacies (Kenny and Mackay 2014).
The Thai Rak Thai, however, by itself represented a new and unprecedented political force in the Thai political landscape. Although operating under a relatively new system, with a constitution passed in 1997, the bigger change to Thai politics was, arguably, that the new electoral framework facilitated the entering of Thai Rak Thai on the political stage in 2001, in which it won a landslide election. The second election in 2005 – the one analysed here – was a tremendous success for the Thai Rak Thai, and Thailand got its first ever single-party government. The Thai Rak Thai, under the leadership of the businessman-turned-politician Thaksin Shinawatra, had managed to mobilize large poorer segments of the Thai population. This mobilization was enabled by a mix of new and old strategies: universal policies introducing cheap health care and micro-credit loans were introduced and hugely contributed to the political success – but in parallel with these new policies, old strategies of working through clientelist networks and vote-buying had to be maintained. The Thai Rak Thai portrayed itself as a new type of Thai party and there was thus a strong emphasis on ‘newcomers’ in general, although there was little specific talk about the inclusion of women. Despite the strong emphasis on a new type of candidate, more representative of the electorate at large, we do not see any effects on the gendered composition of Thai Rak Thai candidates. The representation of women has been consistently low in the Thai parliament, at about 10 per cent (Bjarnegård Reference Bjarnegård2013).
These are two parties that operated in very different settings and political cultures, but they are comparable in the sense that change was on the agenda in both parties. We thus see two cases with windows of opportunities for change, and it is in light of these opportunity structures that we can study strategies used to preserve the status quo (cf. Bergqvist et al. Reference Bergqvist, Bjarnegård and Zetterberg2013). Both parties also had to search actively for new candidates and thus did not have to be as concerned with incumbents as parties usually are. Yet, while ‘newness’ is often considered to be conducive to gender equality (see Mackay Reference Mackay2014), neither of the parties have lived up to expectations on women’s representation. In the case of the Thai Rak Thai, discussions about selecting a ‘new’ type of politician did not address issues of women or gender, meaning that, despite the party’s strong performance in 2001 and 2005, the party did not manage to increase the number of female candidates substantially. Instead of inventing new ways of identifying candidates, the Thai Rak Thai invented refined ways of enticing politicians from other parties to join it. Meanwhile, although the parliamentary ‘face’ of Scottish Labour is female, reflected in its high proportion of women parliamentarians, gendered patterns of turnover within the party have resulted in a decline in the selection and election of female candidates since 1999, suggesting that gender parity and quota mechanisms have been poorly institutionalized within the party (Kenny and Mackay 2014). Thus, rather than invent entirely ‘new’ patterns of selection and recruitment, both parties have, to some extent, fallen back on familiar formulas. They have been unable to distance themselves from the political culture in which they operated and, perhaps more surprisingly, they were unable to free themselves from the stickiness of the informal institutional framework that regulates how candidates are selected.
Two key themes emerged as we compared our findings from the two studies – informal recruitment and informal decentralization. We will now briefly elaborate on these two themes in order to explore their gendered impact and illustrate the value-added of the comparative study of candidate selection.
Formal Regulations and Informal Recruitment
Our case studies shed light on the ways in which party demand is not simply formulated in formal rules but also shaped by informal encouragement and gatekeeping practices. Both parties had formal rules in place, but these rules were either not very specific or not enforced, leaving large room for leeway, interpretations and loopholes ‘on the ground’.
In the case of the Scottish Labour Party, there were extensive formal party regulations for the candidate selection process, including both Scottish and National Executive Committee guidelines for parliamentary selections, and a Candidates’ Code of Conduct. Labour’s initial candidate selection reforms in 1999 were aimed at creating a more fair and open process, intending to reform what had been a relatively closed process of local nomination and selection by ‘unrepresentative, largely male, constituency activists more on the basis of patronage than competence’ (Bradbury et al. Reference Bradbury, Mitchell, Bennie and Denver2000: 151–2). As such, in the run-up to the 1999 elections, the party implemented a number of formal rule changes, including a central panel of pre-approved candidates. It also attempted to professionalize the application process, introducing a person specification, job description and a standard application form. In practice, however, there has been increasing slippage between the formal rules of the recruitment process and their actual enactment and enforcement on the ground after 1999. Formally, for example, job descriptions and person specifications are still in place. But at the constituency level in the run-up to the 2007 elections, for instance, these were not used in many cases, despite repeated requests from both candidate applicants and constituency party members. While formally candidate applicants were also required to have already been pre-selected on the Scottish Labour central panel of approved candidates, in practice this rule did not appear to be consistently enforced in all constituencies, and indeed, in some cases candidates were approved after the fact. Formally, there was little evidence of the central Scottish party taking active measures to recruit particular candidates in the run-up to the elections, but most candidate applicants cited informal encouragement from individuals such as party activists and local party members as a key factor influencing their decision to stand for selection.
In Thailand, in contrast, even the formal regulations for candidate selection are rather brief and leave ample room for flexibility. The national legal framework merely stipulated that all political parties had to have internal regulations that specify candidate selection procedures and rules (Organic Act on Political Parties 1998: Section 11). The internal regulations of the Thai Rak Thai did include two sections on candidate selection, but they were very brief. They basically stipulated that the formal selection be made by the Party Executive Committee and that candidates be party members. Because the candidate selection process is relatively unregulated, it comes as no surprise that the Thai Rak Thai candidates interviewed had followed different paths in order to become a candidate. They all had one thing in common, however: they had been encouraged and sometimes talked into running for office by senior party officials or local party strongmen. None of them had stepped forward themselves, as this was deemed inappropriate.Footnote 5 While party officials claimed that all women had to do in order to become selected was to step forward, this was in stark contrast to the actual process described by the (male) candidates interviewed. Instead, people deemed ‘suitable’ for political office were informally invited and encouraged to stand for election. Often these people were not newcomers to the political arena – instead, far from going out to find a new type of person, the Thai Rak Thai worked hard to encourage the most successful established politicians to change sides. It specifically searched for people with a proven track record of winning elections, and its main target group was therefore sitting members of parliament (from other parties).Footnote 6
In the Scottish case, the underlying trend of informalization was compounded by the overall lack of intervention in the process by the central party and the inconsistent and uneven enforcement of formal selection rules by both central and local party officials. In some cases, the practice of rule-breaking appears to have become a ‘rule’ in itself. For example, while the party’s Candidates’ Code of Conduct explicitly prohibits any campaigning until after the short-listing stage, there was a general understanding among the candidates in the constituency under study that there would be canvassing outside the formal rules: ‘the rules were acknowledged and ignored’.Footnote 7 While this decoupling of formal and informal rules was masked by formal stability on the surface, the day-to-day business of candidate selection was largely guided by informal rules. In the Thai case, however, the formal rules are, in themselves very brief. Party selectors do not have to break formal rules, because the rules stipulate so little about what is supposed to take place. They do have to work out informal arrangements to substitute the absence of formal rules, however (Helmke and Levitsky Reference Helmke and Levitsky2004). Thus, while both parties operate in very different settings, in both cases, rules of informal behaviour and recruitment existed in a context of either weak or ineffective formal rules in which non-compliance routinely went unsanctioned. The gendered consequences of party demand cannot be fully comprehended by analysing party documents. Instead, informal networks of encouragement and recruitment often matter a great deal.
Informal Decentralization and Gendered Local Practices
The necessity of taking formal regulations and informal practices into account is also illustrated by the dynamics of local influence over the selection process in both Scotland and Thailand. If we are to determine where candidate selection takes place, it is necessary to investigate where it actually took place, not simply where the regulations say it takes place. In both these cases, the candidate selection process was de facto, although not de jure, decentralized. In practice, this meant that local interests came to play a large and decisive part in the selection process, with gendered consequences. The comparison between these two cases helps us understand why, as highlighted previously, accounts of the relationship between party decentralization and women’s representation have been inconclusive.
While the trend within the British Labour Party since the late 1980s onwards has been one of greater centralization, evidence from Scotland suggests that the party is now characterized by an increasing degree of territorial autonomy. The decentralization of power within the party increased in the aftermath of the 1999 Scottish Parliament elections, where the party’s centralized approach to candidate selection had attracted criticism for ‘imposing’ certain types of candidates on reluctant local constituencies. After 1999, the British Labour Party still retained final authority over the candidate selection process, through the National Executive Committee. In practice, however, Scottish Labour was able to draw up its own selection procedures without the National Executive Committee’s intervention, although these decisions were taken within a wider framework of centrally prescribed principles. Final selection decisions were left up to party members as a whole. Yet while, formally, the central Scottish party still retained primary authority over candidate selection decisions, in the constituency under study, the party appears to have withdrawn from almost any intervention – formal or informal – into the process, signalling a potential return to past practices of decentralized constituency-based selection (cf. Denver Reference Denver1988). In the Scottish case, central intervention at the constituency level was perceived to be highly contentious, particularly in the area of gender balance.Footnote 8 The lack of central party intervention in the process was therefore welcomed as a positive development by constituency members and candidate applicants, but this de facto decentralization left participants in the selection process with considerable leeway to circumvent and subvert formal rules and reforms, and to fall back on ‘familiar formulas’ of informal local patronage, to which we return below.
The Thai Rak Thai was, by many, perceived to be a very centralized party, due to the huge influence of charismatic party leader Thaksin Shinawatra and the weak organization of its branches. The Thai Rak Thai had a strong top in the party leader and a massive base among its supporters, but no strong institutions in between. The party regulations, too, in spite of their brevity, did point to the Party Executive Committee as the formal authority for matters of candidate selection. The de facto process of selecting candidates was, however, decentralized and even localized. In practice, it was often the responsibility of the incumbent or of a local party strongman to find a new candidate.Footnote 9 Sometimes a poll including the names of local politicians, community leaders, businessmen, teachers and other local notables was conducted in order to find out who was popular in the area in question.Footnote 10 Often, however, the new candidate came from a close circle surrounding a local Thai Rak Thai strongman. If it was not a close relative, it was someone from the local canvassing network. Although the Party Executive Committee had veto power, in most cases it rubber-stamped a decision that was taken locally and based on very local concerns.
These de facto decentralized selection procedures also had gendered consequences in both cases. As already highlighted, despite a detailed formal rule-book, the Scottish Labour Party selection process largely operated in accordance with informal rules and shared understandings. Despite the absence of formal job descriptions and person specifications, for example, participants in the selection process at the constituency level highlighted an informally shared understanding of what selectors were looking ‘for’, repeatedly highlighting the importance of being seen as ‘local’.Footnote 11 Establishing ‘localness’, though, was not an objective matter of residence, but rather relied on informal networks of local patronage: political access and opportunities depended largely on who – rather than what – the candidates knew. As Norris and Lovenduski (Reference Norris and Lovenduski1995: 238) highlight, candidate selection by patronage is based on subjective and informal criteria of ‘acceptability’, where the key question for selectors is whether candidate aspirants are ‘one of us’. Decisions are often made by a limited number of actors, who are usually predominantly male (Matland Reference Matland2005). In the Scottish case study, male participants repeatedly highlighted the importance of ‘playing the game’, being ‘well-connected’, ‘local politicking’ and ‘knowing the right people’.Footnote 12 When interviewees were asked to clarify, the ‘right people’ were usually identified as key local and central party men.Footnote 13 ‘Localness’ also played out in internal party debates over gender quotas. As several constituency members explained, the ‘problem’ with centrally enforced quota mechanisms was not about women candidates specifically; rather, the problem was the central imposition of ‘outsiders’. The repeated linking of gender quotas with ‘imposed central intervention’ positioned female candidates as perpetual outsiders to the process, marking women as ‘Other’. The constructed dichotomy of locals vs. others also disadvantaged particular political masculinities, positioning certain male candidate applicants as ‘outsiders’. Some saw this as part of the strategic machinations of particular local party men, aimed at keeping certain male candidate applicants off the shortlist.Footnote 14 But both male and female outsider candidates perceived this tension in gendered terms, as ‘favourite son’ selections.Footnote 15 Thus, despite attempts to broaden the process of candidate selection after devolution, the evidence suggests that there has been a drift back to the gendered model of the ideal candidate, the ‘local man’ (cf. Lovenduski and Norris Reference Lovenduski and Norris1989).
Similarly, the Thai political landscape is still clearly marked by patronage and clientelism, particularly in rural areas. ClientelismFootnote 16 is an informal institution that requires the building and maintenance of close-knit personal informal networks. Just as in Scotland, then, knowing the right local people and being part of the right local networks is crucial for social acceptance and for being considered a suitable candidate. A strong clientelist network has become close to an informal prerequisite for being a successful candidate in Thailand, and these networks are also the main recruiting grounds for new candidates. What is more, the clientelist networks are highly gendered. They are almost entirely male-dominated and, when asked, male politicians say they want to maximize their chances of electoral success by recruiting people who are already in politically strategic positions in the local area and who have access to localized resources to be distributed to voters. In addition, politicians claim they need to feel that they can trust these people with secretive tasks such as distribution of clientelist goods and money. As in the Scottish case, recruitment was never explicitly framed as a gendered practice, it was more a question of being an outsider or an insider. Being an insider means being someone with access to local funds, important contacts in the local area and with large networks that can be used in the political campaign. Equally important, however, is that the person in question is perceived as someone who can be trusted. Trust is often homosocial, in the sense that we often perceive that we can trust people of the same sex as ourselves. We tend to trust people whom we perceive we can predict, and prediction is easier when we think we see ourselves in other people (Bjarnegård Reference Bjarnegård2013; Collinson and Hearn Reference Collinson and Hearn2005). Male recruiters in the Thai Rak Thai therefore tended to bias their selection in favour of other men. For male recruiters, this implies that other men are seen as more competent and trustworthy and, by default, as insiders and as ideal candidates. One seemingly secure way of choosing an insider is to select a close relative or a son – or someone who is like a son. Everyone wants to be close to and similar to the candidate. Sometimes they even call him ‘father’. The inclusion of a father and son into a network is generally greatly encouraged and this type of relationship is even simulated where no biological relationship exists. A father–son relationship is seen as increasing stability and predictability, as a son is perceived to be similar to his father, or even the same as his father. Thai women cannot approximate the favoured son, and whereas close relationships between two men were seen as stabilizing, close relationships between a man and a woman in the clientelist networks are seen as endangering the predictability and stability of the network, as well as introducing distrust and new types of problems, including sexual relations. Women thus do not have access to the all-important ‘homosocial capital’ that Thai politicians rely on in order to build clientelist networks, make political careers and gain electoral power.Footnote 17
In summary, in both cases localized selection processes created an uneven playing field in which key party actors in positions of power were charged with making, interpreting and enforcing the rules – networks that, in both contexts, continue to be dominated by men. And by virtue of their positional power, these local ‘insiders’ were able to break the rules, or create their own set of rules, using informal and shared understandings to their advantage and keeping outsiders out of the loop (an omission that could then be attributed to the outsiders’ lack of local connections). Thus, in both the Thai Rak Thai and the Scottish Labour Party, there remains a gendered process of boundary construction which privileges certain (informal) institutional interconnections over others. And while the tension between locals and outsiders was presented in gender-neutral terms, at the same time, this constructed dichotomy was profoundly gendered.
CONCLUSIONS
The Scottish Labour Party and the Thai Rak Thai are different parties that operate in different contexts, yet analysis of the gendered aspects of their candidate selection processes suggests that they are also marked by some striking similarities. In both cases, the formal rules of recruitment were either weak or ineffectually enforced, leaving considerable room for actors on the ground to ‘fill the gap’ with informal rules and implicit understandings of ‘how things are done’. In addition, while both parties were (formally) highly centralized, in practice the de facto process of selecting candidates was quite decentralized and even localized. In both cases, we find evidence that localized processes – marked by informal practices of local patronage and clientelism – operate differently for men and women, with women positioned as gendered ‘outsiders’ to the process and therefore unable to gain access to political power.
While revisiting both of these cases points to some fruitful avenues for future study, there are still a number of theoretical and methodological challenges remaining. Questions can be raised, for example, about the comparability of these two settings. However, despite our expectations about the impact of the volatility of the Thai political system, we see that both parties, in fact, attempted to keep to informal and well-known (for insiders) processes for selecting candidates. Outsiders were perceived as less predictable and trustworthy in both environments. Certainly, there may be even more at stake in the Thai candidate recruitment, where the threat of system breakdown is always present, but it is interesting to note that the logics of gendered inclusion are relatively similar in two seemingly dissimilar cases, suggesting that common causal mechanisms may be at play.
An additional challenge, as previously highlighted, is that while institutions have distinctly gendered cultures and are involved in active and ongoing processes of producing and reproducing gender, no institution does this in exactly the same way. This is particularly the case for informal institutions, which are highly contextual, raising questions as to whether the similarities between our case studies are largely coincidental. Nevertheless, comparisons across cases can help to develop at least limited generalizations which may ‘travel’ well across different settings (cf. Pierson Reference Pierson2004). Our cases point to particular gendered mechanisms of institutional resistance and reproduction which may have portability in other contexts. For example, both cases highlight the ‘stickiness’ of informal institutions, drawing attention to how ‘old’ ways of doing things have been reinvented and redeployed, even in new settings – namely informal and masculinist party practices of local patronage, clientelism and homosociality. Certainly there are parallels here with other studies of gender and political institutions, which point to the ways in which male-dominated political elites have shifted the locus of power from formal to informal mechanisms in order to counteract women’s increased access and presence in formal decision-making sites (see, for example, Hawkesworth Reference Hawkesworth2005; Kathlene Reference Kathlene1995; Puwar Reference Puwar2004). Studies of recruitment in other organizational settings also point to homosocial patterns of recruitment and even to the emphasis on favoured ‘sons’ (see, for example, Holgersson Reference Holgersson2013). Our analysis also lends further weight to existing research in the field, which suggests that decentralized candidate selection processes may have negative effects for women, highlighting the gendered dimensions of local interests and influences over the recruitment process, as well as their gendered effects.
Future work in the field therefore needs to explore further the internal party dynamic, while remaining attentive to the gendered and informal dimensions of the candidate selection process. Such a process requires an emphasis on empirical complexity – candidate selection processes operate within a broader institutional and political context and are subject to different spatial and temporal constraints. Untangling the interplay between formal and informal rules and gender in the candidate selection process will require more comparative research across space and across time, as well as more in-depth case studies that situate their findings in relation to the findings of other cases. Given the difficulties of obtaining reliable information on the formal and informal dimensions of the candidate selection process – particularly the need for in-depth empirical research and country- and party-specific expertise – we would suggest that other researchers follow our lead and begin to carry out collaborative research.
Acknowledgements
The authors thank Tània Verge, Joni Lovenduski, panel participants at the 2014 MPSA Annual Conference, workshop participants at the 2014 ECPR Joint Sessions and the 5th LIWEPOP Workshop at the University of Leicester, and the anonymous reviewers of this journal for their helpful comments and suggestions on this paper. Elin Bjarnegård wishes to acknowledge the support of the Swedish Research Council (grant number 421-2010-1638). Meryl Kenny wishes to acknowledge the support of the UK Economic and Social Research Council (PTA-026-27-2122) and the UNSW Vice-Chancellor’s Postdoctoral Fellowship Scheme (RG104436). The authors are equal contributors to this article: names are listed in alphabetical order.
