Convergence versus Divergence of Mass-Elite Political Cleavages: Conceptual, Methodological, and Theoretical Innovations

1. Research Background and Key Questions Two contemporary trends make this workshop timely: 1) diverging political cleavages between the elites and masses, and 2) the increasing influence of non-socioeconomic cleavages in party politics. These trends seem to be prevalent in new democracies. For instance, the East Asian party politics literature has suggested that major politicized cleavages in South Korea, Japan, and Taiwan have revolved around nationalism and foreign policy issues rather than social-distribution ones (McAllister, 2007; McElwain, 2014; Jun and Hix, 2010). Foreign policy/nationalism divisions have been particularly salient among the political elite. However, their relevance to the mass public has faded in comparison to recent socio-economic needs, driven by precarious employment, an ageing society, and declining fertility rates. Old democracies are not immune to the growing divergence between the elites and masses. Two cases in point are the United Kingdom’s referendum result on European Union membership, and Republican candidate Donald Trump’s victory in the US presidential election. On the former, despite the “Brexit” result, pre-referendum surveys showed that more than 80 percent of Westminster legislators were in favour of remaining in the European Union. On the latter, even if Washington and state elites overwhelmingly disapproved of Donald Trump, he won the election by securing 46.2 percent of the votes. Against this background, the primary scope of the workshop are democracies (both new and old) experiencing long-lasting discrepancies of significant scope between the positions or priorities of the masses and elites on important political dimensions, e.g. more than 20 percent gaps between elites and masses on the abortion issue over a decade. The workshop seeks to answer the following series of questions. i) Why do we observe mass-elite discrepancies in some democracies? - Can this be explained by members of the elites having substantially different socio-economic/demographic backgrounds, or are the elites simply “locked-into” particular experiences/identities they share during the early stage of their political career as a cohort? - Has the mass media become an independent actor shaping political cleavage at either mass or elite level? If so, what is the difference between old and new media? ii) What are the ideal ways to conceptualize and measure mass and elite-level discrepancies over time? - How are political cleavages defined in your country/field? What are the pros and cons of loose/strict conceptualizations of political cleavage? - How can we triangulate/prioritize between various results derived from different datasets used to measure elite and mass cleavages? If certain datasets are lacking, how can we maximize the analytical leverage with available data, and using which empirical techniques? Despite the topic’s relevance to contemporary political events, the party politics literature on political cleavage has reached a plateau due to increasing doubts about the validity of conventional assumptions and lack of clear groundwork on varieties of related concepts and measurements. Considering that a persistent mass-elite discrepancy goes against the very idea of representative democracy and can disillusion voters about politics itself, it is imperative to have a coherent and inclusive framework within which we can understand origin and continuance of this trend. For this goal, the workshop here aims to combine insights from two classical schools of party politics—the strategic, temporal, and agent-based Rikerian approach, and the social, historical, and structure-based Rokkanian approach. The next section will flesh out two key questions, by demonstrating the relevant literature, limitations, and gaps the workshop is going to fill. 2. Relevant Literature, Limitations, and Potential Improvements Question 1: Why do we observe mass-elite cleavage differences in some democracies? Lipset and Rokkan Lipset and Rokkan’s (1967) seminal work on cleavage and party politics points out that wealth redistribution is a long-standing source of political disagreement between key parties in old democracies. Their conclusion is supported by the evidence of consistent class-based voting patterns through the late 1970s (Franklin et al., 1992). Moreover, two assumptions that often have been taken for granted in Lipset and Rokkan’s paradigm are that: a) parties are connected through programmatic party competition centred on socio-economic issues; and b) political cleavage formation is a bottom-up process reflecting societal structure and voters’ attributes (Sartori, 1968; Chandra, 2004). However, recent findings show that these cannot be the default assumptions in new democracies, and are increasingly problematic even in old democracies. Recognition of Multiple Political Cleavages and Top-Down Cleavage Formation Process First, the assumed primacy of socio-economic cleavages is challenged by the experiences of non-Anglo-European democracies. The major axis of party competition revolves around ethnicity and regionalism in many African countries, religion and identity in Middle Eastern states, and religion, ethnicity, and caste in Southeast Asia. Indeed, party cleavages have deviated from the traditional Left-Right in some Anglo-European countries such as Ireland, Hungary, or Poland. Second, recent works also demonstrate that cleavages are not simple reflections of structural socio-economic or ethno-linguistic divisions formed in a bottom-up manner. Research in new democracies, such as in Latin America and post-communist Europe, highlight the power of political elites in defining dimensions of competition (Mainwaring and Torcal, 2003) and the role of epistemic communities in creating and maintaining barriers between political camps (Poguntke 2002). Similarly, the elite manipulation of political cleavages is evident in old democracies. For instance, elites can define political divisions by creating identities, forming communities, and reinforcing the structural and cultural distinctiveness of professional groups and classes (Sartori, 1969; Bartolini, 2000; Kriesi, 1998). Possibility of Long-lasting Elite-Mass Discrepancy Despite the development in the literature, what is lacking is the recognition of a systematic distinction between elite and mass. That is, elected politicians and voters can have different priorities and can also be divided along different political dimension for a substantial amount of time. However, the literature tends to equate the two levels as one. For instance, the classical definition both implicitly and explicitly assume that, for any political division to be considered a “cleavage”, it should be observed both at mass and elite level at the same time (Lipset and Rokkan, 1967; Deegan-Krause, 2007). Moreover, mass preferences based on voting patterns or public opinion polls have often been used as a tool for convergent validity testing with measures used to capture elite preferences such as expert surveys—or vice versa. This line of approach treats elite and mass political cleavage as differing sides of the same coin, and is ill-equipped to deal with the discrepancy between the two in both old and new democracies. However, in reality, no matter which direction an attempt at politicization begins—from either the bottom or top—the degree to which a key political cleavage reflects either elite or mass divisions—for either a short or long period of time—should be an empirical question. Potential Reasons behind Long-lasting Discrepancy: Origins and Continuity Why do we observe mass-elite discrepancy? The following four factors can serve as reasonable points for departure. They can be divided into reasons giving rise to the discrepancy in the first place, and reasons related to maintaining an existing discrepancy. First, critical junctures and socialization of elites. It has often been noted that the nature of their formative environment leaves a long-lasting “genetic imprint” on parties (Slater and Wong, 2013; Randall and Svasand, 2007; Panebianco 1988; Riedl 2016). This phenomenon can happen through political elites’ socialization process, as those who went through critical junctures like democratization share particular norms and identities, which in turn, shape the organizational DNA of mainstream parties. The tendency can be reinforced in places where key bureaucrats and politicians are highly homogeneous, coming from similar educational, family, or career backgrounds. This modifies Lipset and Rokkan’s (1967) well-known “freezing cleavage” hypothesis by putting forward the idea that critical junctures have a disproportionate influence in locking-in elite (compared to mass) cleavages. Secondly, the party politics literature has not paid due attention to the influence of new actors beyond mass and elite that can shape/mediate political cleavages. The news media is an excellent example in this regard. The media is often thought as a linkage between the mass and elite as it conveys key pieces of information to form one’s opinion. Although several works have shown the partisan bias of the media mirroring that of the elite and mass, not much work has shown the independent/mediating influence of media in forming political cleavages at multiple-levels. However, with the Brexit case, one can emphasize the role of the media by pointing out that the division between key newspapers mirrored that between the mass and the elite—The Daily Express and The Sun on the mass side, and The Times and The Guardian on the elite side (Berry et al., 2015; Bruzelius, 2014). Another potential influence that often goes unnoticed can be found the dividing line between old and new media. Old media like newspapers or television news disproportionally reflect mainstream elite views, while new media like Facebook or Twitter accommodate and reinforce contrasting views of the masses. Salient examples include anti-establishment/populistic leaders declaring war against the old media, e.g. Trump in the US, Erdogan in Turkey, Orban in Hungary. Third, electoral rules, party systems, constitutional structure, election campaign laws, or candidate selection process can be particularly important in perpetuating the existing cleavage discrepancy. To begin with, a clear example can be found in countries with two-party systems. Usually coinciding with a first-past-the-post electoral rule, in such systems, large parties receive more seats than their vote share warrants, so two large parties tend to be “manufactured”. This can then suppress small parties from emerging by promoting potentially important issue resonating with the public. The marginalization of pro-Brexit party UKIP is a clear illustration of this—in the 2015 U.K. general election, it received 12.6 percent of the total votes but won only one seat (0.2 percent of the total seats). Fourth, among new democracies, the key party-voter linkage is often known be to be based on charismatic leadership or clientelistic quid-pro-quo exchanges (Kitshelt, 2015). Going beyond new democracies, evidence increasingly suggests the “programmatic party-voter” linkage cannot be taken as the default assumption even for more matured developed democracies. For instance, as for the clientelistic linkage, several works have repeatedly shown that the clientelistic linkage between voters and politicians often adapts and endures in the face of economic development (Hicken, 2011) and democratic transition (Kitschelt, 2000). The fact that clientelistic party-voter linkage often cross-cuts existing political cleavages or reflects needs of a small proportion of the population is enough to cause lasting mass-elite discrepancy. Question 2: What are the most ideal ways to conceptualize and measure mass and elite-level differences over time? Political Cleavage as concept: Ideal Conceptual Boundary and Qualitative Differences In relation to concepts, Rokkan (1970) did not clearly define political cleavage itself. Even though a large number of studies use this concept (more than 5,000 Google-scholar citations on Lipset and Rokkan’s works related to political cleavage), a uniform definition of “cleavage” is still lacking (Bertoa, 2014). On the one hand, many scholars intuitively sense that the concept should not be about just any political attitude, behaviour, or conflict at a specific time. On the other hand, we should also be careful not to define the concept too narrowly and limit its analytical value. For instance, making encompassing interests, normative or attitudinal outlooks, and a strong organizational base (Bartolini, 2000) as necessary conditions of “cleavage” can be too strict a definition, since there can be stable and consistent political divisions in a country without self-conscious demographic groups with distinct political organizations that share a similar mind-set. For instance, “issue cleavages” such as between material and post-material values (Inglehart, 1997) or foreign policy orientations (Lijphart, 1984) often cross-cut socio-demographic boundaries without having a distinguishable organizational structure. And these issue divides are increasingly relevant, since their effect on voting compensates for the declining role of structural cleavages (Dalton, 1996). The core concern of this project is that elites and masses can be divided on different dimensions and form political cleavages. In light of this, the workshop strikes a balance between strict and loose conceptualizations, and defines political cleavage as “a long-lasting political division,” while discrepancy indicates that “division” on a particular political dimension is present only at mass or elite level for a substantial period of time, e.g. more than two to three electoral cycles. However, the workshop will endeavour to find trade-offs between looser and stricter concepts in light of criteria of good conceptualization (e.g. resonance, coherence, familiarity, differentiation) in established social science scholarship (Gerring, 1999; Sartori, 1984). Given the global scope of this workshop and the existence of a wide range of key political cleavages across regions, more academic attention needs to be devoted to qualitative differences between various political cleavages. Particularly noteworthy dimensions include: i) differing degrees of visibility, e.g. easily observable structural cleavages (religion, income, ethnicity, race, language, or location) versus difficult-to-measure value cleavages (post-material values, foreign policy orientations, national identity, or tension between democratic and authoritarian values); ii) different degrees of closure, e.g. more static and exclusive cleavages like gender and race versus more fluid cleavages like age and income; and iii) between cleavages with different degrees of consensus-reaching possibility, e.g. identity issues versus social distribution issues (Evans and Whitefield, 1998). Measuring Cleavages with Multiple Datasets and Methods Considering that this workshop’s central goal is to understand mass-elite discrepancies over time in both new and old democracies, it is imperative to triangulate between various datasets using different methods. At the national level, the most straightforward way of measuring mass-elite discrepancy on a particular political dimension is to compare the (elite) roll-call vote result with (mass) referendum result (provided that the same voting question is used which captures the concerned cleavage dimension). Simple as it may be, these results are rare and, if they exist at all, yield no more than one or two time points. Lack of these data notwithstanding, recent years have seen an explosion of newly accessible information on the web, such as floor speeches or debates in the legislative arena (Grimmer and Stewart, 2013) and digitized newspaper articles on political issues (Bornschier, 2010), along with the maturation of conventional datasets such as long-standing panel studies and cross-country attitudinal surveys. Beyond this, the weaknesses of lacking reliable and valid cross-national data in measuring political cleavages can even be complemented using analytical equivalent context-specific indicators (Adcock and Collier, 2001). Reflecting this trend, Shim and Plsek (2018)’s meta-analysis on empirical works related to political cleavage shows that more than 70 different datasets were used for 112 analyses. Starting from “comparability of findings” between empirical works, the abundance of data sources creates new lines of challenges which need to be addressed. We ought to be particularly careful in choosing and combining specific datasets by considering the motivations behind data generation, the extent of coverage (e.g. to what extent particular dataset includes various issues and diverse political actors), continuity (e.g. how frequently the dataset is generated), and latent dimensions (e.g. whether the data captures policy-oriented issue positions or vote-oriented issue saliency). With these potential obstacles in mind, the workshop will shed new light on i) how to avoid inappropriate datasets/minimize downsides of each dataset or measurement techniques (Laver, 2014); ii) how to specify, extract, organize, and rescale results derived from heterogeneous sources collected by multiple organizations with different purposes (Schedler, 2012); and iv) importance of collective research efforts in both data collection and data application. 3. Pursued paper types, likely participants, and expected output The workshop seeks papers that improve our insights into elite and mass cleavage politics over time, as well as necessary groundwork on related concepts and measurements. For this goal, the scope of analysis is limited to places where significant discrepancies between elites and masses can be observed on important political dimensions for a substantial stretch of time. Although the key focus of “discrepancy” is on political division (measured from positional stances), the workshop is also interested in examining “discrepancy” on political saliency (measured from issue priorities). Papers can be either theoretical or empirical—or combinations thereof—directly addressing one of the two specified questions or other important aspects of long-lasting mass-elite discrepancies not pointed out in this proposal, such as i) is mass-elite discrepancy always negative? or ii) what are the impacts of mass-elite discrepancies on voter turnout, party position identifications, or quality of policy commitments? Empirical papers can be single-case studies with a comparative approach, as well as small-or large-N analysis. Particularly encouraged are papers with an inter-regional comparative outlook including new democracies in Africa, Asia, and post-communist Europe. This workshop will include both junior and senior scholars from the disciplines of comparative politics, political sociology, or political concepts and multi-methods, who are interested in political dynamics of cleavages at multiple levels. This proposal has been endorsed by two ECPR standing groups—“Political Sociology” and “Elites and Political Leadership”. Several colleagues had already expressed their interest on this theme, and a mini-conference on mass-elite discrepancy was held in June 2018 at the German Institute of Global and Area Studies (Hamburg, Germany). As a result, we are now in the process of registering the topic as an ECPR Research Network, and at the same time, building a related database (“political cleavages around the world”) based on secondary literature and expert interviews. Upon successful completion of the workshop, we plan to organize an edited volume (or two) with some of the qualified papers selected for the workshop. With this possibility in mind, scholars who already have a working paper or at least some relevant analyses are encouraged to apply.