Ruling the Void by Peter. Mair

(2002: 168) has usefully pointed out that the very logic of the ‘parties-matter’ thesis builds from two core propositions: first, that the ‘social constituencies of political parties in constitutional democracies have
distinctive preferences
and successfully feed theprocess of policy formation with these preferences’; and second, that the ‘policy orientations of political parties broadly
mirror
the preferences of their
social constituencies’
(see also Keman, 2002). It follows that in the absence of such constituencies there is little by way of collective preferences that can be mirrored, even if the parties could or wished to mirror them, and hence the whole logic of the partyness of policy-making becomes difficult to sustain.
DECLINING ELECTORAL COHESION
    It is beyond dispute that the once-distinct electorates of the mainstream political parties in western Europe have become markedly less cohesive in the past two to three decades. To be sure, it can be shown that traditional cleavages remain relevant to voting behaviour. For all the changes that have been wrought in the economy and in the polity over the past decades, for example, workers are still more likely than the middle class to vote for left-of-centre parties, and active church attenders are still more likely than secular voters to support religious parties. This is undeniable (e.g., Elff, 2007). But what is also clear is that the relative weight of these voting determinants has declined. Church attenders might still vote along religious lines, but there are many fewer such citizens in European electorates than there were thirty years ago, and their capacity to shape electoral politics has diminished accordingly (Best, 2011). The shifts in class voting are even more marked. The core working-class constituencies have experienced pronounced demographic decline, while the homogeneity of political preferences within the remaining class cohorts has been lost. In the most comprehensive and nuanced comparative study to date, Oddbjørn Knutsen (2006) points toa substantial decline, both absolute and relative, in class voting in western Europe since the mid-1970s, with the falls being most pronounced in precisely those polities where class had once been a very strong predictor of political preference (see also Knutsen, 2007).
    It is also beyond dispute that, in responding to, and sometimes even provoking, the changes in their electoral alignments, parties have become electorally more catch-all, easing their grip on once-core social constituencies while extending their appeal ever more broadly across traditional class and religious lines. In part, of course, this is the inevitable result of social change. Since the core constituencies themselves have begun to decline or to fragment, there is less within the social structure for the parties to grip on to (see also Freire, 2006). Voters have become more ‘particularized’ (Franklin et al., 1992). But in coming to terms with this social change, the individual parties have also had to learn to be more attractive to those segments of the electorate once seen as beyond the pale: religious parties have had to learn to appeal to secular voters, socialist parties to middle-class voters, liberal parties to working-class voters, and farmers parties to urban voters. In other words, it is not only that the vote has become more free-floating and available: so also have the parties themselves, with the result that political competition has come to be characterized by the contestation of socially inclusive appeals in search of support from socially amorphous electorates.
    The tendency towards the decline of collective identities within western electorates, resulting from more or less common socio-economic or socio-cultural processes, has therefore been further accentuated at the political level by the behaviour and strategies of the competing political parties, and one consequence of this has been to undermine the foundations of