Cleavage Politics in the 21st Century: How and why social background affects political preferences

Abstract

Existing scholarship shows that social background – represented by features such as class, religion, or gender – is an important predictor of political preferences, including party preference, across Western democracies. This relationship has recently come to the attention of a broader audience, with the referendum on Brexit and the election of president Trump. Unfortunately, we know much less about why social background relates to political preferences, i.e., which mechanisms induce the relationship? The main motivation for this thesis is to contribute to filling this knowledge gap by answering the overarching research question: How and why does social background affect political preferences? The four papers that constitute this thesis study the question through the application of the same theoretical framework, namely cleavage theory. The first paper (chapter 2 in this thesis), which is co-authored with Oddbjørn Knutsen, focuses on a measurement issue that is important for the chapters to follow. The paper takes on the prominent claim that changes in the employment structure in post-industrial societies require a redrawing of the class map, especially to capture new distinctions within the middle class. We find that the traditional EGP class schema predicts modern political preferences equally well as the new schema endorsed by critics of the former schema, suggesting that the worries about the continued relevance of the traditional EGP class schema seem exaggerated. This justifies its use in later chapters. In papers 2 and 3 (chapters 3 and 4), I empirically analyse the importance of one of the most important propositions in cleavage voting theory, namely the idea that social groups have indirect effects on party preference through political values. I find that political values account for less than half of the variation in both class voting and religious voting, meaning that we need to look for additional explanations in order to satisfactorily understand the phenomenon. I also find that there is a substantial direct effect, due to frozen group-party alignments and group identity. Furthermore, the share of the effect that is direct or indirect varies substantially between parties and party families. The two papers jointly develop a theory of cleavage voting. The crux of the theory is that the effect of social group membership on party preference is indirect through political values to the extent that social groups hold distinct sets of political values, and to the extent that political parties emphasise issues of relevance to those values. This means that party characteristics affect not only the level of cleavage voting, but even the very mechanisms behind such voting. While I test – and find support for – the theory on the two most studied cleavages, namely class and religion, I propose that it may be valid for any kind of cleavage voting. Finally, the fourth paper (chapter 5), which is co-authored with Rune Stubager, scrutinises another prominent claim, namely that globalisation has transformed, surpassed, or even replaced the old cleavages such as class or education. Our findings suggest that direct measures of exposure to globalisation have limited independent effects on both core political attitudes and party preference. Furthermore, there is no indication that class and education today primarily works through globalisation-related processes, nor that these processes interact with class to create cross-class alliances. In conclusion, social structural variables still affect party preference, and while the dominant view has been that this is primarily an indirect effect through political values, it seems that direct effects related to group identities and frozen party alignments play a larger role than commonly assumed. While globalisation processes as of yet have not significantly altered the way social structural variables affect political preferences, it seems that political actors themselves are able to affect which mechanisms that are at work.