Ruling the Void (5 page)

Whatever the reasons for any fall in levels of participation, therefore, these analyses seem mainly quite sanguine about the trends. Long-term stability in levels of participation has been followed by a slight decline, but this is not so great that it need be a source of worry for those concerned with the healthy functioning of modern democratic life. Is this a reasonable conclusion? On the face of it, and especially with regard to the European data, the interpretation is certainly plausible.
⁴
Thus, through each of the four decades from the 1950s to the 1980s, average turnout levels in western Europe scarcely altered, increasing marginally from 84.3 per
cent in the 1950s to 84.9 per cent in the 1960s, and then falling slightly to 83.9 per cent in the 1970s and to 81.7 per cent in the 1980s. This was essentially the steady-state period, as has been emphasized by Norris and Franklin. That said, the decline from the 1970s to the 1980s, while small, was remarkably consistent across fifteen long-established European democracies, with just three countering an otherwise general trend: in Belgium, where voting is compulsory, turnout increased slightly from 92.9 to 93.9 per cent from the 1970s to the 1980s; in Norway, where turnout increased from 81.6 to 83.1 per cent; and in the Netherlands, where mean turnout remained more or less unchanged. In each of the other twelve countries for which long-term data are available, however, mean levels did in fact decline in the 1980s, whether marginally, as in Austria, which recorded a fall of less than 1 per cent, or more substantially, as in France, which recorded a fall of more than 10 per cent. The decline may have been marginal when looked at cross-nationally, but it was almost universal, and hence might well have justified some concern.

More important, it is a trend that began to accelerate in the 1990s and beyond, with average turnout across western Europe falling from 81.7 per cent to 77.6 per cent in the last decade of the twentieth century, and to 75.8 in the first decade of the new century. To be sure, even at this level, which is the lowest recorded in any of the postwar decades, turnout remained relatively high, with an average of slightly more than three-quarters of national electorates casting a ballot in the elections held during the 1990s, a figure that remains substantially higher than that recorded in nationwide elections in the United States, for example. Even allowing for this, however, and for the fact that the drop from the 1980s to the 2000s is little more than 6 per cent, it is nevertheless striking to see the overall European figure
in the 1990s dipping below the 80 per cent level for the first time in five decades. Here also, moreover, there is a striking consistency across countries, in that, looking back from the turn of the century, eleven of the fifteen democracies involved also recorded their lowest ever decade averages in those ten years. The exceptions to this pattern again include Belgium, where the decade averages are almost invariant, but where the lowest level was recorded in the 1960s, and Denmark and Sweden, which both recorded their lowest levels in the 1950s. Even in these three cases, however, it should be noted that the average level of turnout in the 1990s was lower than in the 1980s. The fourth exception is the United Kingdom, which was unusual in recording its trough in participation in the 1980s. Indeed, the UK is the only one of these fifteen countries that recorded even a marginally higher level of turnout in the 1990s than in the 1980s, although in this case turnout later plunged to a remarkable low of just 59 per cent in the first election of the twenty-first century.

This pattern is therefore very striking, and all the more so when account is taken of the sheer extent of the decline in particular countries. In Austria, for example, where turnout had remained safely above the 90 per cent level in each of the preceding four decades, the drop in the 1990s was almost 8 per cent. Similarly sharp declines were recorded in Finland, in Germany, which had absorbed the new voters of the former Democratic Republic during this period, and in the Netherlands and Norway. Even more striking, although across the longer term, is the case of Switzerland, where the then exclusively male electorate recorded an average of 69 per cent turnout in the 1950s – higher than that recorded in France or Ireland during the 1990s – but which, this time with equal rights for women, recorded an average of less than 44 per cent in the 1990s. In other words,
more and more countries experienced record low decade averages in the 1990s, these in some cases reflecting very sharp declines.

This trend has persisted into the twenty-first century. As noted, the 2001 election in the UK was marked by the lowest level of turnout since the advent of mass democracy. The 2002 parliamentary elections in both France and Ireland were also marked by historic low levels of turnout, and while Ireland picked up again in 2007, France fell to a new record low of 60.4 per cent. Record lows were also recorded in 2008 in Italy and 2001 in Norway, as well as in 2002 in Portugal, and in 2000 in Spain. Levels close to historic lows were recorded in Greece in 2000, in Switzerland in 2003, in Austria in 2006 and in Finland in 2007. Why the trend towards ever lower levels of participation has continued remains, of course, an open question, to which we will return. It may simply reflect generational shifts. It may also be because of sheer boredom. The key point, however, is that we are seeing something that is both unidirectional and pervasive, and that offers a striking indicator of the growing enfeeblement of the electoral process.

There is one other way of seeing this picture that is perhaps even more telling. Indicators of turnout change are somewhat like those of climate change: the shifts we see do not necessarily occur in great leaps or bounds, and are not always linear. Moreover, while indicating withdrawal and disengagement, change in turnout levels is often registered as a trickle rather than a flood. For these reasons, and again like the indicators of climate change, the importance of what is often just a slight or uneven trend may be underestimated or even disputed. One way in which climatologists get around this problem is by laying less stress on the trends as such, and by drawing attention instead to the patterns that are visible in the timing and frequency of the peak values. This is, in fact,
a very simple approach to measurement, and is also intuitively meaningful. Thus, for example, in a publication from 2003, Phil Jones and Anders Moberg adduced clear evidence of global warming by noting that the warmest decade on record had been the most recent, the 1990s, while 1998 emerged as the warmest single year, followed by 2001. Further evidence of global warming was adduced by noting that the eight warmest years on record had all occurred since 1990, even though in that same period air temperatures were also recorded (for example, in 1992, 1993 and 1994) that were little higher than those reached in the late 1970s. In other words, the pattern is evident, even if the trend is not wholly unidirectional.

The same is true for turnout levels, and indeed for many other indicators of mass political behaviour, and for this reason the sheer extent of change at this level is also often underestimated. Although there is no undisturbed downward trend in levels of electoral participation, for example, record lows now come with greater frequency, and in a greater number of polities. As can be seen from
Table 1
(a) overleaf, which lists the three elections with the lowest levels of turnout in each of the fifteen long-established European democracies, almost four-fifths of these elections have taken place since 1990. In other words, not only do the last decades hold the record for the lowest turnout of any postwar decade in western Europe, but within the great majority of west European democracies, most, or even all of the record low turnouts have occurred since 1990. The two clearest exceptions are Denmark and Sweden, where, seemingly for unremarkable contingent reasons, the lowest-turnout elections fell in the 1950s. Beyond these cases, the only other odd exceptions are one low-turnout election in the 1960s (in Belgium), another one in the 1970s (again in Belgium), and two in the 1980s (in France and Luxembourg). The remaining thirty-five cases all date from 1990 or later. In other words, however small the overall shifts in turnout might be, they are nevertheless clustering together in a remarkable fashion – see
Table 1
(b). Indeed, this pattern also extends to the newer southern European democracies: the three lowest levels of turnout recorded in post-authoritarian Greece were those in 1996, 2000 and 2007; in Portugal, the lowest levels were recorded in 1999, 2002 and 2005; and in Spain in two of the three lowest turnouts fell in 1989 and 2000 (the third was in 1979). Here, as in the long-established democracies, the more recent the
elections, the more likely they are to record troughs in participation. There is no certainty here, of course; like the pattern evinced by climate change, turnout sometimes bucks the overall trend, even today. However, the overall direction and reach of the change is unmistakable, and it offers the first strong indicator of the increase in popular withdrawal and disengagement from conventional politics.
⁵

The second key aggregate indicator relates to the behaviour of those citizens who do participate, and measures the extent to which their voting patterns reveal consistency and stability over time in the distribution of partisan preferences. Those citizens who continue to vote in elections are clearly still engaged with conventional politics, however marginally. As popular involvement fades, however, and as indifference grows, we can expect that even these citizens who do continue to participate will prove more volatile, more uncertain and more random in their expressions of preference. If politics no longer counts for so much, then not only should the willingness to vote begin to falter; so also should the sense of commitment among those who continue to take part. Choices are likely to prove more fickle, and to be more susceptible to the play of short-term factors. In practice, this also means that election outcomes are likely to prove less and less predictable. Electoral volatility is likely to increase; new parties and or new candidates are likely to prove more successful; and traditional alignments are likely to come under pressure. Hand in hand with indifference goes inconsistency.

As with patterns in turnout, expectations of growing unpredictability in the balance of party support in national party systems in western Europe have been current for a number of years. Here too, however, the empirical record at the aggregate level has usually failed to confirm them. Thus while party systems in some countries did indeed experience a substantial increase in their levels of electoral flux through the 1970s and 1980s, others appeared to become even more settled than before, resulting in what was generally a ‘stable’ and relatively subdued level of aggregate electoral change across western Europe as a whole (Bartolini and Mair, 1990). For many observers, such findings proved puzzling, since the evidence from survey data in particular had suggested that in the 1970s the western democracies had already begun to experience symptoms of breakdown in their traditional electoral alignments and historic cleavage voting patterns (Dalton et al., 1983; Franklin et al., 1992). As it turned out, however, these undeniable changes at the level of individual behaviour did not seem to translate into equivalent shifts in the party balance at the aggregate level. Indeed, even by the end of the 1980s, aggregate electoral volatility on a European-wide basis remained relatively muted, while many of the traditional parties that had already dominated electoral competition in the 1950s or even earlier continued to be serious contenders. These older parties had certainly seen some of their aggregate support slipping away to the benefit of new formations, but even by the end of the 1980s it was striking to see how much of their overall vote share they managed to retain.

This is borne out by the mean levels of aggregate electoral volatility in the period from 1950s to the 1980s. The measure applied here is that originally proposed by Mogens Pedersen (1979), who calculated the level of volatility simply by summing the (aggregate) electoral gains
of all winning parties in a given election, or, which is the same thing, the (aggregate) electoral losses of all losing parties. It is, of course, a crude aggregate measure, and it may well underestimate the real level of vote switching – as measured by individual survey evidence, or whatever. As an aggregate measure, however, it has the advantage of being calculable for all elections, including those in the distant past as well as those in polities where survey data are either absent or unreliable. In any case, the point here is to note that by this measure, contrary to many expectations, levels of aggregate electoral volatility across the fifteen long-standing democracies in Europe scarcely changed between the 1950s and the 1980s: the west European national average fell from 7.9 per cent in the 1950s to 6.9 per cent in the 1960s, and then rose to just 8.9 per cent in the 1970s and in the 1980s. This was hardly the stuff of electoral earthquakes. That said, these averages did conceal quite a bit of flux within the individual party systems. Denmark, the Netherlands and Norway moved from remarkably quiet elections in the 1950s to relatively unstable elections in the 1970s, before returning again to more stable outcomes in the 1980s. In contrast, both France and Germany began the postwar period marked by the substantial flux of postwar political reconstruction, and then settled down to more steady-state politics in the 1960s and 1970s. In other words, while the average level of aggregate electoral volatility in western Europe as a whole tended to remain quite stable, this was partly masking the contradictory patterns in the experiences of the different polities.