Ruling the Void (6 page)

Here again, however, as with the evidence of turnout, the more important observation is that this picture began to change significantly in the 1990s. Across western Europe as a whole, the 1990s became the peak decade for electoral volatility, with a score of 12.6 per
cent, almost 4 points higher than that recorded in the 1970s and 1980s. Not too much should be made of this, of course. On a scale with a theoretical range running from 0 to 100, and which even here had a range of decade averages that ran in practice from 2.5 (1950s Switzerland) to 22.9 (1990s Italy), a mean value of 12.6 still reflects more (short-term) stability than change. On the other hand, the 1990s was the first of the five postwar decades in which the overall mean of instability breached 10 per cent, and also the first decade to record such a major shift from the previous mean value.

The significance of the 1990s is also underlined by reference to the individual national experiences. In all but four of the countries (the exceptions are Denmark, France, Germany and Luxembourg), the 1990s constituted a new national peak in volatility levels, which, in the majority of cases, easily exceeded 10 per cent: 10.8 in Belgium, 11.0 in Finland, 13.7 in Iceland, 11.7 in Ireland, 22.9 in Italy, 19.1 in the Netherlands, 15.8 in Norway and 13.8 in Sweden. Indeed, of all the individual national elections held in the 1990s, close to two-thirds recorded volatility levels in above 10 per cent. This confluence was also unprecedented, and again signalled that the patterns at the end of the century were markedly different from those of the earlier postwar years.

Moreover, and as with the turnout data, there was little sign that these new excesses had begun to abate in the new century. In elections in 2002, both Austria and the Netherlands experienced record high levels of aggregate instability, as did Italy in 2001. France, Norway and Sweden also recorded remarkably high levels of volatility in their first twenty-first-century elections, although in these cases no absolute records were broken. More generally, as can be seen in
Table 2
, a clear majority of the most unstable national elections to be recorded since 1950 have occurred since 1990.

The very simple approach to presenting the data here is again borrowed from the climatologists, and follows the breakdown applied to the turnout data in
Table 1
. In this case the pattern is not so one-sided: volatility data inevitably prove more erratic than turnout data, being quickly responsive to political crises as well as to institutional and social-structural change (Bartolini and Mair, 1990: 253–308). Nevertheless, the period since 1990 again proves exceptional: not only do 60 per cent of the record national highs in volatility fall in this period – one-third in the 1990s, more than a quarter in the next ten years – but it is also noteworthy that no other decade
comes close to matching this clustering. Indeed, in no earlier decade does the number of high volatility elections even come close to double figures. Once again, the more recent the elections, the less likely they are to yield a predictable outcome.

Since 1990, in short, ever fewer voters have seemed ready to participate in elections, although turnout levels in themselves have remained reasonably high, while among those who do participate, there has been a greater likelihood that they will switch their preferences from one election to the next.
⁶
Not only has each of these indicators reached a relative extreme in the period since 1990 (whether recording troughs in turnout, or peaks in volatility) across western Europe as a whole, they also tend to the extreme in a large majority of the individual polities. That is,
both
extreme lows in turnout
and
extreme peaks in volatility have been recorded since 1990 in almost all of the long-established European democracies. The exceptions were Luxembourg, which had very low turnout but only moderate volatility; Sweden, which recorded high volatility but not exceptionally low turnout; and Denmark, which proved extreme on neither indicator during this recent period. Beyond these cases, the evidence of unusual patterns since 1990 is not only striking, but also remarkably consistent. Across western Europe, citizens, where they are not abstaining from the ballot altogether, are voting with significantly reduced partisan commitment. In these heightened levels of instability, we see a second strong aggregate indicator of disengagement.

The same message comes through more and more clearly from individual-level survey data. That is, the often substantial shifts evinced by these aggregate data on turnout and volatility now correspond closely to the evidence of individual-level experiences as tapped by election studies and commercial polling projects. Many of these latter data have been collated and summarized by Dalton and Wattenberg in their comprehensive
Parties Without Partisans
(2000), and what is also striking in this instance is both the consistency and the pervasiveness of the various changes that have been observed. One key indicator is the degree to which individual voters feel a sense of belonging or commitment to particular political parties, a feeling captured by various measures of partisan identification. And here (
Table 3
) decline is very evident: in eleven of the thirteen countries,
including a number of non–European Union polities, for which relevant data are available – the exceptions are Belgium and Denmark – the percentage of voters claiming a sense of identification with parties has fallen over the past two decades or so. Even more strikingly, the smaller numbers of voters who report a strong sense of belonging or identification has also decidedly fallen, and this time in every single one of the countries concerned. As Dalton notes, it is not just the scale of the decline that is important here, but also the fact that it occurs in all of the cases for which data are available. There therefore seems little that is either contingent or circumstantial, and as with so many other data discussed in this chapter, ‘the similarity in trends for so many nations forces us to look beyond specific and idiosyncratic explanations … For public opinion trends to be so consistent across so many nations, something broader and deeper must be occurring’ (Dalton, 2000: 29).

Table 3
Trends in party identification in western Europe, 1960s–1990s

Yet more evidence of this broader and deeper process can be seen in the other sets of survey data that Dalton and his colleagues marshal. Split-ticket voting, for example, whereby voters opt for one party in one electoral arena, and for another party in another electoral arena, has risen in all those cases where it can be measured over time (Australia, Canada, Germany, Sweden and the United States). A committed and engaged voter, with strong partisan loyalty, will normally vote for the same party regardless of the arena involved – for example, voting Democrat in US presidential and congressional contests, as well as probably in local state and county elections. Lesser partisan commitment and lesser engagement are more likely to be associated with more free-range voting patterns, and hence with a greater willingness to split the ticket, and it is this latter practice that is growing. Voters are also less ready or less able to decide in advance how they will vote, preferring to observe the campaign, or
even remaining uninterested, until closer to polling day. Here too, with a single Danish exception, this pattern is more and more prevalent, with almost every election study reporting a substantial increase in the proportion of voters who make their decision how to vote either during the campaign or only shortly before polling day. Again, the implication is a lack of stable commitment on the part of voters, and hence also a lack of engagement. It is also hardly surprising, then, to see that these voters are also far less likely to engage in more demanding campaign activities, whether this might be by way of attending political meetings, working for a party or candidate, persuading others to vote in a particular way, or even donating money. On almost all of these measures, and in almost all the countries for which data are available, the survey evidence once again clearly points to decline: individual voters are less and less willing to participate in this more demanding sense – for many, at least as far as conventional politics is concerned, it is enough to be simply spectators.

Citizens are also obviously much less willing to take on the obligations and commitments associated with membership in party organizations. Here too, it is striking to note not only the sheer decline in the number of party members over time, but also the extent to which this decline seems characteristic of all long-established democracies (Van Biezen et al., 2009). Although the pattern here is more pronounced than in the case of changes in levels of turnout or of electoral volatility, the conclusions that have been drawn about party membership levels tend to reiterate in an even more compelling way those drawn about the more conventional levels
of participation. In other words, through to the 1980s, the evidence of decline in this form of political engagement was somewhat ambiguous and disputable. From the 1990s onwards, by contrast, the trend has been unequivocal and seemingly unstoppable.

The first major study based on aggregate – often official party – data was that summarized in Katz and Mair (1992), and covered a large number of European polities from the beginning of the 1960s through to the end of the 1980s. This study found that, while there was a decline in the numbers of party members as a proportion of their respective national electorates (the only exceptions were the cases of Belgium and West Germany), which were themselves expanding substantially in the population booms of this period, there was little evidence of decline in the actual numbers themselves. Indeed, by the end of the 1980s, it appeared that party membership had actually grown in absolute terms not only in Belgium and West Germany, but also in Sweden, Norway and Italy, while falling both in absolute terms and as a share of the electorate in Finland, the Netherlands, Austria, Denmark and the UK. A mixed picture, in other words, and one that led at the time to the conclusion that, contrary to prevailing expectations, there had been no wholesale ‘collapse’ in membership (Katz and Mair, 1992: 332). This conclusion was supported by a study that incorporated much of the available individual-level survey evidence on the topic (Widfeldt, 1995), and later again by a comparison of World Values Study survey data from the early 1980s and early 1990s. Indeed, far from seeing the party membership ratio entering a general decline through to the early 1990s, this latter analysis suggested that it had actually grown, and sometimes quite substantially, in such countries as Finland, Iceland, the Netherlands, Belgium, Norway, Britain and France. There seemed little evidence from these particular data
that these countries were experiencing ‘a spreading disillusionment with partisan politics’ (Norris, 2002: 134, 135). On the contrary, the picture was one of vibrancy and engagement.

This relatively sanguine picture was shattered by the end of the 1990s. By then, and regardless of whatever conclusions might have been drawn from the survey data, the patterns in the aggregate data had become unequivocal. The first real check of the new patterns was reported in Mair and van Biezen (2001), and included data on thirteen long-established European democracies, as well as a number of the newer democracies. In each of the established cases, it turned out that the ratio of party membership to the electorate at large had fallen markedly between the beginning of the 1980s and the end of the 1990s (see also Scarrow [in Dalton], 2002: 86–95). That is, in not one of these cases had the membership ratio remained steady, let alone increased. The summary figures were truly striking: In 1980, an average of 9.8 per cent of the electorates in these thirteen long-established democracies were party members; by the end of the 1990s, the figure had fallen to just 5.7 per cent. To put it another way, and to trace the contrast back even further, at the beginning of the 1960s there were ten democracies in Europe for which it is possible to trace reliable membership figures, and across all ten the average membership ratio was 14 per cent; in a majority – six out of ten – of the countries, the ratio was above 10 per cent. In other words, more than one in every ten eligible voters were members of political parties. At the end of the 1990s, by contrast, there were twenty countries for which it was possible to find reliable membership data, some old democracies, some new. Across all twenty, the average membership ratio was just 5 per cent, little more than a third of the level recorded in the early 1960s, and of these twenty
countries, only one – Austria – recorded a ratio exceeding 10 per cent.
⁷