Curtice J. (2005) 'Turnout: electors stay home again' Parliamentary Affairs 776.One feature of the outcome of the 2001 election sent shockwaves through politicians of all parties. Just 59.1% of the electorate in Great Britain voted.1 Not only did this represent a drop of 12.4% on the turnout in 1997, it was the lowest recorded level of voting participation since 1918. Given that many troops were still abroad in that post-First World War election, it is probably the case that more of the electorate abstained voluntarily in 2001 than in any previous UK election since the advent of the mass franchise. It appeared that the British political class had become seriously disconnected from the public it sought to serve, thereby raising some important questions about the effectiveness of British democracy.
Turnout was little better in 2005. Just 61.2% of the electorate in Great Britain voted, an increase of no more than 2.1% on the record low of four years previously. At least things had not become any worse, but turnout was still well below the norm in British general elections. Between 1970 and 1997 turnout had never dipped below 71.5%, and it had even been as high as 79.1% in February 1974. In other words, turnout in 2005 was still about ten points below what had hitherto been its low water mark.
Moreover, turnout was once again so low despite the fact that it was possible for any elector to vote by post in the comfort of their own home. No longer was access to a postal vote confined, as it had been previously, to those who were ill, away on business, or on holiday. The new more liberal regime had been in force at the time of the 2001 election, but as the new regulations had only been introduced some five months previously, there had been relatively little opportunity for people to become aware and to take advantage of the new facility. So while the number of postal voters doubled in 2001, the proportion of the electorate registered to vote by post was still no more than 4%. It might be anticipated that postal voting on demand would have had rather greater impact in 2005.
This chapter therefore examines two main issues. First, why was there only a modest increase in turnout compared with 2001? Does it mean that British democracy does indeed face some kind of 'crisis' of disconnection between its politicians and the public? Secondly, what impact did the availability of postal voting on demand have in 2005? How far did people avail themselves of the facility, and what contribution, if any, did it make to the increase in turnout which occurred?
Turnout before 2005There are two main kinds of explanation for declining turnout. One account argues that voters are no longer as strongly motivated to go to the polling station as they once were. Fewer have a strong emotional attachment to a political party.2 More are cynical about politics and distrustful of politicians.3 A better educated electorate finds that putting a cross on the ballot box is an inadequate means of expressing its political views and no longer feels the obligation to do so. And a more affluent public with a diversity of leisure opportunities apparently has less time for, or interest in, politics. An alternative explanation, however, argues that the key to the low turnout in 2001 was a failure by parties to mobilise voters.4 The election offered the electorate little choice. First, fewer people than ever before thought there was a great deal of difference between the Conservative and Labour parties. Meanwhile, apart from one very brief spell during the 'fuel crisis' of September 2000, the Labour party had enjoyed an unprecedented near ten year period of continuous double digit leads in the opinion polls. Voters were told that the outcome of the election was a foregone conclusion, and in any event they felt there was not much difference between the principal contenders for power anyway.5 In those circumstances, while those citizens who did have a strong motivation to participate largely still turned out, those with a weaker commitment to voting or little interest in politics, the kind of elector who always has to be mobilised to go to the polls, stayed at home. The pattern of turnout in mid-term elections held during the course of the last parliament offered some support to this latter interpretation. By 2003, Labour's dominance of the electoral landscape was under challenge. Even if the ideological distance between Labour and the Conservatives remained narrow, other political parties such as the United Kingdom Independence Party, the British National Party, and Respect were offering distinct platforms on issues with substantial emotional resonance, such as Europe, immigration, and Iraq. In these circumstances, turnout showed signs of recovery. Turnout in five English by-elections held in hitherto safe Labour seats - but ones where in many cases Labour now faced a serious challenge from the Liberal Democrats-averaged 40.1%, well above the 24.9% recorded in by-elections held in such seats in the 1997-2001 parliament and even the 37.8% in the 1992-97 parliament. A record high turnout of 38.2% was recorded in the 2004 European elections. While this figure was boosted by the use of all-postal ballots in some areas and by holding the election on the same day as local elections in others, it still appears that turnout would have been close to the level of the 1979 and 1984 European elections even if those measures had not been in place.6 In any event, turnout in the contemporaneous 2004 local elections was above the norm for such elections in recent years.7
Turnout in 2005
After these indications that perhaps the British electorate might be returning to the ballot box, albeit in mid-term elections that usually fail to command the participation of as many as half of voters anyway, the low turnout in the 2005 election could be regarded as something of a disappointment. Perhaps voters have lost the motivation to vote after all? Or can we demonstrate that the election might still have seemed a rather dull one to many voters?
One piece of evidence collected during the election campaign suggests that voters were indeed no less motivated in 2005 than they had been previously. During the middle of the election campaign MORI asked about how much interest people have in politics, a question that it has administered on a number of occasions over the last thirty years. As Table 1 shows, in fact slightly more people than ever before said they were 'very' or 'fairly' interested in politics. As many as 61% fell into one or the other of these two categories, more than matching the previous all-time high of 60% recorded in 1973 and 1991. On this evidence, at least, it is difficult to argue that the low turnout in 2005 is a reflection of any marked reduction in the willingness of voters to participate.
Note: The 2001 figure is for the whole of the United Kingdom; the remainder are for Great Britain only.
Sources: 1973-2004; Extracted from Electoral Commission, An Audit of Political Engagement 2, Electoral Commission Hansard Society, 2005; MORI/Financial Times poll April 2005.
So were voters once again simply not motivated by parties to vote? Certainly, as Table 2 shows, it remained the case that relatively few felt that there was a great deal of difference between the parties. A NOP poll conducted just before the election campaign in 2005 found that only 21% believed that there was a 'great deal' of difference between the Conservative and Labour parties, up only slightly on the record low of 17% in 2001, and still well short of the previous low before that of 33%. Moreover, those who said that there was a great deal of difference between the parties were 15 points more likely to say they were absolutely certain to vote than were those who did not see such a large difference between the parties.
2. Indicators of Potential Mobilisation 1945-2005
na: not available.
Sources: A. Heath and B. Taylor, 'New Sources of Abstention?' in G. Evans and P. Norris (eds), Critical Elections: British Parties and Voters in Long-Term Perspective, Sage, 1999; C. Bromley and J. Curtice, 'Where Have All the Voters Gone?' in A. Park, J. Curtice, K. Thomson, L. Jarvis and C. Bromley (eds), British Social Attitudes: The 19th Report. Sage, 2002; NOP/Independent poll February 2005.On the other hand the opinion polls suggested a much closer contest than they did in either 1997 or 2001. No longer was the message one of double digit leads but rather of a more modest lead that averaged just five points in the final polls. While this still meant that the contest looked less close than the majority of elections held in the 1950s and 1960s, at least the election no longer appeared to be a walkover. Thanks to the pro-Labour bias in the electoral system,8 however, when the findings of the polls were projected into seats, they were typically represented as pointing to a third three-figure Labour majority in a row. The prospect of a Conservative victory appeared to be utterly remote. Thus arguably the message of the polls was still one that suggested little doubt about the outcome.9
So it appears that while citizens might have seen slightly more reason to vote than they did in 2001-a somewhat closer contest and perhaps a slightly bigger gap between the two main parties-for the most part the stimulus to vote was probably still relatively weak in 2005, and it was certainly insufficient to generate anything more than the most modest of increases in turnout. But if the closeness of an election makes a difference to voters, then we should find that not only does it affect the total number of people who vote but also where people vote. Those living in marginal constituencies would appear to have more reason to participate than those who live in safe constituencies. Parties have more incentive to campaign heavily in these areas too.Table 3 indicates that those living in marginal constituencies were indeed more likely to vote. On average the turnout was nine points higher in the most marginal constituencies (where the lead of the winning party over the second party was less than 5% in 2001) than it was in the safest constituencies (where the winning party had a lead of 20% or more). Furthermore, despite the fact that many of the marginal constituencies had also been close prior to the 2001 election and had therefore experienced below-average drops in turnout between 1997 and 2001,10 the increase in turnout was also nearly one-point greater in the most marginal seats than it was in the safest.
Source: Author's calculations from checked and corrected version of BBC Election Results database.
It is not clear though that we can ascribe all of these differences between marginal and safe seats to the closeness of the contest. For Table 3 also indicates that the picture in safe Conservative seats is very different from that in Labour ones. Turnout was, at 65.2% , well above the national average in safe Conservative seats while the increase in turnout, +2.3%, was also a little above average too. Neither of these figures quite matches that in the most marginal Conservative seats, but the gap in both cases is only of the order of one point. It was exclusively in safe Labour seats that turnout was particularly low and increased less compared with 2001.This suggests that perhaps it is something about the type of people living in safe Labour constituencies that also helps to account for the low turnout in these seats, and not simply the apparent inevitability of the local outcome. One possibility is that these seats include a relatively large proportion of disaffected Labour voters who are persistently and increasingly opting to stay at home-the gap between the turnout in safe Labour seats and that in the rest of the country has grown at each and every election since 1992.11 However, there is no consistent evidence in this election that Labour's vote fell most where turnout fell most. But, equally as likely, it may also be the case that those living in more socially deprived neighbourhoods, neighbourhoods that are most commonly found in safe Labour constituencies, may have been particularly affected by the lack of stimulus to vote at the last two general elections. Certainly those in working class occupations, those who have few educational qualifications, and those who are unemployed are all more likely to evince relatively low levels of interest in politics, and are less likely to feel they have a duty to vote, the very characteristics that we saw earlier were associated with falling levels of turnout in 2001.
These alternative possibilities are difficult to disentangle. But multivariate analysis indicates that even after we take into account Labour's share of the vote in a constituency, there remains a relationship between both the level of turnout in 2005 and the change in the level of turnout since 2001, and indicators of the social characteristics of a constituency, such as the percentage who say they are in good health or the percentage who are routine manual workers. We illustrate this in Table 4 which shows that even if we confine our attention to safe Labour constituencies, turnout was both higher and increased more in seats with relatively high proportions of people in good health and with relatively few routine manual workers. So it may well be the case that the reason why turnout is so low in safe Labour seats is not simply because they are safe but also because they contain a disproportionate number of people who need to be mobilised to go to the polls, and for whom nether the local nor the national political contest provided much incentive to vote in 2005.
Note: Based only on safe Labour constituencies, defined as those won by Labour in 2001 with majorities of 20% or more. Percent in good health: Low-Less than 65% of adults in good health (2001 Census); High-more than 65% in good health. Percent routine manual workers: Low-Less than 11% of those aged 16-74 employed in routine manual occupations (2001 Census); High-more than 11%.
Source: Author's calculations from checked and corrected version of BBC Election Results database.
The 2005 election appears then to have been similar to 2001 in its failure to provide voters with a stimulus to vote. As in 2001, few saw much difference between the parties, while in practice the message of the opinion polls was still that only one party appeared to have a chance of winning. This lack of mobilisation may have had a particularly strong impact in safe Labour constituencies, not just because they are safe, but also because voters in such seats are particularly likely to require some stimulus to vote. Certainly, as in 2001, there is little evidence that the low turnout is simply or even primarily an indication of a new disinclination on the part of voters to vote.
But perhaps even the two-point increase in turnout overestimates the degree to which there was any greater interest in this election than there was in its predecessor. Maybe the increase is simply accounted for by the easier availability of postal voting. After all, being able to vote at home and then to put one's ballot in the nearest post box certainly reduces the time and effort involved in voting, and any reduction in the 'costs' of voting should make it more likely that people actually cast a ballot.
Full details of how many people voted by post are not available at the time of writing; however, we do have access to information collected by the BBC on the number of postal votes issued for 594 of the 627 constituencies in Great Britain where an election was held. Meanwhile, by modelling these data we have been able to impute the likely figures in the remaining 33 constituencies. We also have information collected by the Electoral Commission for each constituency on the number of postal votes issued and the number validly returned at the last election. This means we can also examine how the increase in postal voting varied from constituency.As we anticipated, there was a substantial increase in the number of people opting to vote by post. Just under 12% of the electorate in Great Britain were issued with a postal vote, treble the proportion who were issued with such a vote in 2001. Given that those who opt to vote by post are more likely to actually cast a vote, this suggests that around 15% of all valid votes cast were cast by post.12 Just as we might expect turnout to be higher in marginal constituencies, so we might imagine that postal voting would be higher in such seats too. Voters would appear to have more incentive to ensure that they did not miss out on the chance to vote should they be unable to reach the polling station on election day. More importantly, perhaps, political parties, aware that those registered for a postal ballot are more likely to vote, have a particular incentive to encourage their supporters to apply for a postal ballot. Indeed the role of political parties in encouraging and enabling people to apply for postal votes became the subject of some controversy after the 2004 local and European elections, most notably following a well publicised case of electoral fraud in Birmingham, and this has since led to the government proposing that political parties should not handle such applications on behalf of voters.13
There is, however, little evidence that postal voting was more common in marginal constituencies. The proportion of the electorate issued with a postal vote in the average marginal constituency (one with a 2001 majority of five points or less) was 11.8%, slightly below the equivalent figure, 12.0%, for the average safe seat (one with a 2001 majority of 20 points or more). Meanwhile the latter figure for safe seats also represented a rather larger increase on 2001 (+8.3%) than the former figure for marginal seats (+7.6%). If political parties were attempting to encourage people to vote by post in marginal seats they appear, collectively at least, to have been remarkably unsuccessful.What does appear to have made a difference to the incidence of postal voting is familiarity. Applications to vote by post were notably higher in those parts of the country with previous experience of voting by post. This can be seen in Table 5 which shows the average proportion of the electorate registered to vote by post in each of the eleven government regions, and the increase this represented on 2001. Postal voting was by far the highest and increased most in the North East. This was not only one of the regions which had undertaken an all-postal ballot in both the 2004 Euro elections, but it also conducted a further such ballot in the regional referendum that autumn, and it was an area in which a relatively large proportion of local authorities had undertaken all-postal ballots in local elections prior to 2004. Requests for postal ballots were second highest in another region where an all postal ballot was held in 2004, Yorkshire & Humberside. Although the increase in postal voting was lower in the North West, the third region to have an all-postal ballot in 2004, than it was in the East Midlands, the latter is a region where local authorities appear to be particularly keen to promote postal voting.14 The Electoral Commission has now lost its former enthusiasm for all-postal ballots on the grounds that the system denies voters the opportunity to choose how they would like to vote. Nevertheless it appears that electors, encouraged perhaps by local authority administrators, are in fact more likely to opt to vote by post once they have had experience of doing so.
Source: Author's calculations from checked and corrected version of BBC Election Results database.
But did this increased take-up of the facility to vote by post have any impact on the overall level of turnout? Did the ease of voting by post mean that those who might not otherwise have voted did so? Or was it simply the case that those who chose to vote by post were for the most part the relatively interested and engaged who would have voted anyway?
If the increased use of postal voting helped to increase turnout then we should find that turnout rose most in those constituencies where postal voting increased most. This is the analysis performed in Table 6, which shows how much turnout increased since 2001 according to how much the proportion of the electorate issued with a postal vote increased.
Source: Author's calculations from checked and corrected version of BBC Election Results database.This analysis suggests that the wider availability of postal voting had, at most, a small impact on turnout in the 2005 election. Turnout did rise least in those seats with the lowest increase in postal voting over the previous four years and most in those seats where the increase was biggest. But the difference between them was little more than half a point. Making it easier for people to vote by post may give people the right to choose how they vote but it evidently does not, unlike all postal ballots,15 make them significantly more likely actually to cast a ballot. Meanwhile the fact that turnout increased by at least one and a half points in those constituencies which had had only modest increases in postal voting certainly suggests that the overall increase in turnout since 2001 cannot simply be accounted for by the wider availability of postal voting.
In the event the political circumstances in which the 2005 election was fought were too similar to that of 2001 to engender anything more than a modest increase on the record low turnout of that year. Voters were as interested in politics as ever, but they continued to feel there was little choice between the parties, while the opinion polls appeared to point once again to a comfortable if somewhat narrower Labour victory. Lacking any stimulus to vote, many again stayed at home. Not even the prospect of being able to avoid the journey to the polling station enticed many voters to exercise their franchise. Turnout depends not on giving people a choice about how to vote but rather on what they are voting about.
But does this mean that we can be sanguine about the prospects of turnout eventually returning to its previous levels once political circumstances change? Not necessarily. One long run change that affects voters' likelihood of going to the polls has occurred over the last forty years: a decline in party identification. Now in more recent years sections of the British electorate have begun to lose the habit of voting. Indeed in the case of some younger voters, they have not had the chance to develop the habit in the first place. Together these developments probably mean that the degree of stimulus required to produce a turnout of 70% or more is now rather greater than it was ten years ago. Unless that stimulus comes soon, perhaps too many people will have lost the habit of voting entirely.
1 All references in this chapter are to Great Britain excluding Northern Ireland, unless otherwise stated. The recent history of turnout in Northern Ireland has been very different from that in the rest of Great Britain. Turnout actually rose slightly to 68.0% in 2001, when the election took place in the midst of a major debate about the future of devolution in the province. In 2005, however, by which time there seemed to be no immediate prospect of restoring devolved political institutions, turnout fell by 5.1 points to 62.9%, the lowest ever in a Westminster election since partition in 1922.2 I. Crewe and K. Thomson, 'Party Loyalties: Dealignment or Realignment?' in G. Evans and P. Norris (eds), CriticalElections: British Parties and Voters in Long-Term Perspective, Sage, 1999. 3 P. Norris (ed.), Critical Citizens: Global Support for DemocraticGovernance, Oxford University Press, 1998; C. Bromley, J. Curtice and B. Seyd, 'Political Engagement, Trust and Constitutional Reform' in A. Park, J. Curtice, K. Thomson, L. Jarvis and C. Bromley (eds), British Social Attitudes: The 18th Report-PublicPolicy, Social Ties, Sage, 2004; R. Dalton, Democratic Challenges,Democratic Choices, Oxford University Press, 2004. 4 C. Bromley and J. Curtice, 'Where Have All the Voters Gone?' in A. Park, J. Curtice, K. Thomson, L. Jarvis and C. Bromley (eds), British Social Attitudes: the 19th Report, Sage, 2002; M. Franklin, Voter Turnout and the Dynamics of ElectoralCompetition in Established Democracies since 1945, Cambridge University Press, 2004. 5 See the argument developed in A. Heath and B. Taylor, 'New Sources of Abstention?' in G. Evans and P. Norris (eds),Critical Elections: British Parties and Voters in Long-TermPerspective, Sage, 1999. 6 J. Curtice, S. Fisher and M. Steed, 'Appendix: Analysis of the Results' in D. Butler and M. Westlake, BritishElections and European Politics 2004, Palgrave, 2004. 7 C. Rallings and M. Thrasher, Local Elections Handbook 2004, Local Government Chronicle Elections Centre, 2004. 8 R.J. Johnston, C.J. Pattie, D.F.L. Dorling and D.J. Rossiter,From Votes to Seats: The Operation of the UK Electoral Systemsince 1945, Manchester University Press, 2001. 9 Indeed, a poll conducted just before polling day by Populus for The Times found that no less than 78% thought that Labour would win an overall majority, with nearly one in three of this group anticipating a Labour majority of more than 100. 10 J. Curtice and M. Steed, 'Appendix 2; The Results Analysed' in D. Butler and D. Kavanagh, The British General Election of2001, Palgrave, 2001. 11 J. Curtice and M. Steed, 'Appendix 2; The Results Analysed' in D. Butler and D. Kavanagh, The British General Election of1997, Macmillan, 1997. 12 In the average constituency in 2001, the proportion of votes cast by post was 1.286 times the proportion of voters issued with a ballot paper. 13 A number of the allegations about the possible abuse of postal voting, including the case in Birmingham, have involved areas with high Asian populations. On average, however, the proportion of the electorate registered to vote by post was not particularly high in constituencies with relatively high Asian populations. On average, in those constituencies where according to the 2001 Census more than 10% of adults are from an Asian background, just over 11% were registered to vote by post. For details of the government's proposals on postal voting, see Department for Constitutional Affairs, Electoral Administration: A PolicyPaper for Discussion, Department for Constitutional Affairs, 2005; Electoral Commission, Securing the Vote, Electoral Commission, 2005. 14 Electoral Commission, Electoral Pilots at the June 2004 Elections, Electoral Commission, 2003. 15 J. Curtice, S. Fisher and M. Steed, 'Appendix: Analysis of the Results' in D. Butler and M. Westlake, BritishElections and European Politics 2004, Palgrave, 2004.