7 VOTING BEHAVIOUR 1
Michael Laver 1
Voting turnout 2
Candidates versus parties 8
Determinants of party support 10
Policy positions of voters 18
References and further reading 24
People writing about Irish voting behaviour have often presented Irish politics as peculiar when set in a European context – a system of ‘politics without social bases’ (Whyte, 1974; Carty, 1981). Certainly, the differences between Fianna Fáil and Fine Gael, combined with the very modest long-term performance of the Labour Party, are not easy things for a specialist in Irish politics to explain to even the most sophisticated of European visitors (see chapter 5). For the most part it is quite clear that Irish party politics has not been motivated, as it has in many other European counties, by a conflict between blue-collar and white-collar workers over the distribution of wealth. Neither has it been driven forward, as it has in many countries with a substantial Catholic population, by a conflict between pro-clerical Christian democracy and anti-clerical secular liberalism.
These general observations were combined by Richard Sinnott with the findings of earlier researchers, into a comprehensive survey of the opinion poll evidence on the social patterning of Irish voting behaviour. He found that there was indeed some patterning, but that this is rather weak when seen in a wider European context (Sinnott 1995: 181-8; see also the books in the How Ireland Voted series: Laver, Marsh and Sinnott, 1987; Gallagher and Sinnott, 1990; Gallagher and Laver 1993; Marsh and Mitchell, 1999; Gallagher, Marsh and Mitchell, 2003). The main reason for this is the appeal of by far the most successful party in the state, Fianna Fáil, to voters from across a wide range of the social spectrum. Fine Gael and the Progressive Democrats are indeed somewhat more middle class in their support profiles than their main party rivals, but Fianna Fáil’s populist appeal, combined with the resulting weakness of the Labour Party, are essentially what have led people to regard Irish party politics as a system without social bases. Sinnott did nonetheless find opinion poll evidence which indicated that things might be changing. Following a series of bitter referendum campaigns on divorce and abortion, he suggested that there may be signs of an emerging religious-secular cleavage in Irish voting behaviour (Sinnott 1995: 193-195). Following the enormous transformation of Irish society that has resulted from a long trend of substantial migration from more rural to more urban areas, Sinnott’s analyses of constituency variations in actual voting patterns also hinted at an emerging rural urban-rural divide that he felt merited further investigation (Sinnott 1995: 114-144).
Useful though the opinion polls summarised by Sinnott and forming the basis of the received wisdom to date undoubtedly were, they were nonetheless based on small and less-than-perfect samples of Irish voters, who were asked only a limited range of questions on behalf of the media organisations that commissioned them. This situation changed radically in 2002, with the carrying out of the first Irish National Election Study (INES). This was a full-scale academic post-election survey, equivalent to those that have been conducted over a long period in most other western democracies. During the course of the INES, a large sample of randomly selected voters were asked a comprehensive range of questions about their voting habits. Thanks to the INES, therefore, we now know far more than we did before about what makes Irish voters tick. This chapter thus goes beyond most traditional work on Irish voting behaviour and is largely based the new insights offered to us by the INES. In what follows, we will explore three types of question:
why do Irish voters vote at all?
given the single transferable vote (STV) electoral system that requires Irish voters to rank individual candidates rather than political parties, do those people who do vote think more in terms of candidates than of political parties?
to the extent that Irish voters do think in terms of parties, what conditions their choice of party?
The question of why voters bother to turn out to vote at all has stimulated a lively debate within political science. This arises from the fact that is extremely unlikely that the result of any major election will be decided by a single vote. Whether electors realise this or not, it is simple common sense that each individual vote, taken on its own, is almost certain not to make a difference to the eventual outcome of the election. Thus if voting is seen as an ‘instrumental’ act, undertaken to make some difference to the world, and if the act of voting is seen as costly in some sense, then going out and casting a vote does not seem a rational thing to do. Yet in the real world a lot of people do in fact vote. Does this mean that they wildly over-estimate the likely impact of their votes, or are simply irrational? This question is often referred to as the ‘paradox of turnout’ or the ‘calculus of voting’ problem within the professional literature (for a good and accessible review of this see Hinich and Munger, 1997).
An alternative view of the act of voting is that this is not something that people instrumentally engage in so as to achieve some measurable effect, but is rather an act of self-expression that is valued in and for itself. On this account the benefit of voting comes simply from voting itself, from the value that citizens attach to using the ballot box to express their own personal views about politics (for a recent and lively review of this point of view, see Schuessler, 2002). If voting helps to fulfil voters’ desires to express themselves politically, and there is of course no reason to regard people who think like this as either stupid or irrational, then the problem of why people turn out to vote becomes one of understanding which factors cause some people to get more, and others to get less, personal satisfaction from the act of voting.
For whatever reason, the percentage of the electorate turning out to vote in Ireland has been declining steadily over recent decades, a trend that Ireland shares with many other west European countries. Figure 7.1 shows levels of general election turnout in Ireland between the 1973 and 2002 Dáil elections and highlights an inexorable downward trend. This decline in turnout has certainly been viewed within the Irish political establishment as a serious problem and various remedies have been explored – including extending voting hours, holding elections on different days of the week, slightly relaxing provisions for postal voting, and, with a view to assisting functionally illiterate voters, even including pictures of the candidates on the ballot papers.
[figure 7.1 about here]
We get at least something of an idea about what is affecting levels of voter turnout by looking at the very obvious variations of turnout rates between different types of constituency. Lyons and Sinnott (2003) remark on the striking contrast between urban and rural constituencies in the 2002 election (continuing a long trend found in previous elections), with turnout lowest in Dublin South-Central at 51 per cent, and highest in Cork North-West at 72 per cent. Paradoxically, although actually getting to the polling station is often physically easier in urban than in rural areas, there is a striking tendency in Ireland for urban areas, and particularly Dublin, to have lower turnouts. Of course many of the ‘urban’ areas into which so many Irish voters have moved in recent years are in reality part of a huge and anonymous suburban sprawl around greater Dublin, which contrasts strikingly with the deeply rooted social structure of the traditional rural areas that, as we shall see, have proved much more resistant to political change.
Moving beyond patterns in constituency level election results to possible individual motivations for non-voting, Irish National Election Study (INES) interviewed a sample of 2,663 people drawn from the electoral register – not just those who actually voted in the 2002 election. This means that we now have comprehensive information about both voters and non-voters in Ireland. The first thing to note in this regard is that the election study considerably overstates the level of turnout in the 2002 election. Of those who agreed to be interviewed, 81 said they voted in the 2002 election,1 as opposed to the 62 per cent of the electorate that we know actually did vote in reality. Overstating turnout is a general trend in election studies, suggesting either that there is a ‘selection bias’ in these studies that arises because people who do not vote in elections are also much less likely to co-operate with academic surveys, and/or that survey respondents interviewed a few weeks after the election have, shall we say, a somewhat rosy view of what they did in the past. Either way, we might reasonably infer that the pattern of systematically over-reported turnout in election studies suggest that non-voters do not seem particularly proud of their behaviour.
Nonetheless, we can find information in the INES on a substantial number of declared non-voters in the 2002 general election. Indeed, we can go further, since there were in fact three occasions in 2001 and 2002 on which Irish voters were asked to turn out and vote. In addition to the general election, there were referendums on the Nice Treaty in 2001 and on abortion in 2002 (the second Nice referendum, in October 2002, took place after the fieldwork for the survey had been completed). If we add up the number of times that voters claim to have turned out on these three days, this number can be anything from zero to three, giving a more complete picture of which types of people are really committed to voting and which are not. Figure 7.2 shows the impact of gender, age and educational attainment on the propensity of people to vote in modern Ireland. While we must bear in mind the impact of selective recall on absolute levels of reported turnout, differences in turnout rates between different social groups are nonetheless very instructive.
[figure 7.2 about here]
First, from the top panel of figure 7.2, we see that gender does not have a huge impact on turnout rates. Nonetheless, concentrating on hard-core non-voters, we do see that about 14 per cent of male electors claimed they did not vote at all despite the three opportunities offered them during 2001-02, as opposed to about 8 per cent of women. Women were in contrast slightly more likely to have voted twice or all three times, but differences in turnout rates between Irish men and women do not seem to be large.
Second, the middle panel of figure 7.2 shows us that turnout during 2001-02 was greatly affected by the age of the voter – a finding echoing that in most of the other research on the matter (see, for example, Lyons and Sinnott, 2003). The impact of age on turnout rates really is very striking, with about 28 per cent of voters in the 18-25 age group not voting at all in this period, as opposed to about 5 per cent of those aged between 36 and 65.2 In contrast, only about 20 per cent of voters in the 18-25 age group claimed to have voted in all three elections, as opposed to well over 60 per cent of those aged between 56 and 65. Young people in Ireland, as elsewhere, do seem very much less likely to turn out and vote than older electors. With only one election study at our disposal, it is not possible to tell whether this is a ‘lifecycle’ or a ‘cohort’ effect. In other words, it could be that young people tend not to vote, but tend to start voting as they move through their lifecycles and get inexorably older. Or it could be that there is now a generation, or cohort, of young people who tend not to vote, a pattern that will continue to manifest itself even when they get older. Nonetheless, if low turnout is seen as a ‘problem’ in Ireland, as it clearly is by some, the INES makes it clear that it is a problem caused much more by high levels of non-voting among younger electors than among older ones.
Third, and again echoing previous work (Lyons and Sinnott , 2003) the bottom panel of figure 7.2 shows us that educational attainment, and in particular a very low level of educational attainment, is also associated with non-voting. Electors who did not complete their primary education were very much more likely to have voted in none of the three elections or referendums of 2001-02. Beyond this, the patterns are not strong, although those with some post-second level education were distinctly more likely than others to claim to have voted in all three elections. Overall, therefore, hard core non-voting tends to be associated with being male, having low levels of education, and above all with being young. The relationship between income levels and non-voting is not shown in figure 7.2 because this is in fact very weak.
Moving beyond basic social and demographic characteristics, it might be that the likelihood of voting is affected by electors’ attitudes to the political system, in particular by whether people feel that their votes make a difference. We recall that people who expect their vote to make a difference, and who feel that it does not, may not consider it worth the effort of voting. The INES asked two questions directly relevant to this. People were asked to agree or disagree with the statement ‘so many people vote, my vote does not make a difference to who is in government’. A second question referred to making a difference ‘to which candidate is elected’. Strikingly, 69 per cent of respondents (in relation to government) and 72 per cent (in relation to candidates) either disagreed or disagreed strongly with these statements. Most Irish voters do not do the same sums as political scientists; they do feel their votes make a difference to the eventual result. Answers to these two questions can be combined into a single index measuring the extent to which people feel their vote makes a difference and respondents can then be categorised into those who feel that their vote makes a big difference, almost no difference, or something in between.3
The top panel of figure 7.3 shows rates of voting turnout among those that felt their votes had a low, medium or high impact, respectively, on the outcome of the election. As we might expect, rates of voting were much higher among those who felt that their vote did have an impact on the election result, and much lower among those who felt it did not. Rates of hard core non-voting (zero votes in the three elections) were almost six times higher among those who felt their vote had a low impact, than among those who felt it had a high impact. We must of course be alert in this context to a problem of cause and effect. Demographic factors cannot possibly be affected by voting patterns – voting a lot cannot, alas, make you younger or more highly educated. However, voting patterns may either affect, or be affected by, attitudes to voting. Thus it might be that people who think their vote has an impact are more inclined to vote for this reason, or the direction of causality might work the other way. It might be that those who vote a lot justify this to themselves by believing that their vote has an impact. This would be an example of the well-known psychological effect of people reducing the ‘cognitive dissonance’ between apparently contradictory aspects of their social attitudes and behaviour. People might be uncomfortable with the idea that they vote a lot despite the fact that they also believe voting makes no difference, and adjust their beliefs about the impact of their vote accordingly.
[figure 7.3 about here]
A second reason for not voting might be that, whatever difference people believe their vote to make to the result of an Irish election, they believe that Irish election results make little difference to their everyday lives. The INES asked a battery of questions on this matter.4 These questions can be combined into an index measuring what we might think of as respondents’ views on the impact Irish politics has on their world.5 The middle panel of figure 7.3 shows that, once more, the effects were in the expected direction. Those who think that Irish politics does not make much of a difference to what happens were much less likely to vote than those who think it does make a difference. At the other end of the scale, high levels of voting were more likely among those who felt that Irish politics does make a difference.
Finally, it may be that some people feel that they just don’t know about, or understand, politics, and are less likely to vote as a result. The INES asked two questions on this matter. People were asked to agree or disagree with the following statements: ‘sometimes politics and government seem so complicated that a person like me cannot really understand what is going on’; and: ‘I think I am better informed about politics and government than most people’. Again we can combine responses to these two questions into a scale capturing each respondent’s sense of ‘understanding’ of politics.6 The bottom panel of figure 7.3 shows the effects on turnout rates of these feelings about Irish politics. The effects are more modest than those we have just discussed. Nonetheless, they are very much in expected directions. People who feel that they don’t understand politics were more likely to be hard-core non-voters. People who feel they do understand politics are much more likely to be regular voters.
So far, we have looked at all of these effects of turnout one at a time, but obviously all effects are likely to be interrelated. Young or poorly-educated people could feel that they can’t understand politics well, for example, so it is important to get some idea of the impact of each factor on turnout levels, holding other factors constant. This can be achieved by using a particular technique of statistical analysis to predict whether or not respondents will be hard-core non-voters (never voting in 2001-02), and whether or not they will be regular voters (voting three times) on the basis of the variables reported in figures 7.2 and 7.3, holding all other variables, as well as household income, constant.7 The results are not reported in detail here, but can be summarised easily.
Trying to explain hard-core non-voting, the significant effects come from gender, education, age, and how much impact the respondent thinks that his or her vote actually has. Men are almost twice as likely as women to be hard-core non-voters. People who did not complete primary education are over seven times more likely than others to be hard-core non- voters, as are people in the 18-25 age group. People in the 26-35 age group are between two and three times more likely to be hard-core non-voters. Finally, there is a strong and systematic link between feeling that the vote makes a difference to the result and the rate of voting. Turning to the syndrome of regular voting at elections, gender appears to have no effect on this, while the effects of age and education remain very strong. Poorly educated and younger people are very, very much less likely to be regular voters. The strongest predictors of regular voting are having a post-secondary education, being older, and feeling that your vote makes a difference.
Candidates versus parties
It is common to describe Irish politics as being highly ‘personalised’, a product in part of a localised political culture, possibly reinforced by the STV electoral system (see chapters 4 and 9). STV does indeed require voters to rank people, not political parties, when voting. The huge amount of constituency work done by most TDs, and even by aspiring candidates, combined with vigorously personalistic local campaigning and publicity, also mean that voters are strongly appealed to by individual candidates. On the other hand, of course, Irish governments are formed and run by political parties – on their own or in coalition with others. The business of the Dáil is run very much by the party whips. At the level of the national news media, the election campaign is fought out between political parties, and the key issue is often presented in terms of which party leader will become Taoiseach. The question that all of this poses, therefore, concerns whether Irish voters are primarily supporting a candidate or a party when they allocate their first preference vote.
The INES asked whether ‘the party or the candidate’ was ‘more important in deciding how you cast your first preference vote in the general election’, which might seem to settle the matter once and for all. Things are a little more complicated, however, since the larger Irish parties typically run more than one candidate in each constituency. Voters supporting such parties are thus forced to choose between different candidates of the same party when allocating a first preference – in this sense their choice of first preference is inevitably a choice between candidates rather than parties. Their answer to a simple ‘party or candidate’ question is thus ambiguous since those who said that they were choosing a candidate could ‘really’ be party voters allocating first preference votes between different candidates of the same first-choice party. A second INES question comes closer to what we want to know, asking: ‘if this candidate had been running for any of the other parties would you still have given a first preference vote to him/her?’ A respondent who answers ‘yes’ to this question is more unambiguously choosing a candidate rather than a party.
Table 7.1 reports and cross-tabulates answers to these two questions. Looking first along the bottom row we see that 62 per cent of respondents said they were choosing a candidate, while 38 said there were choosing a party – this on the face of things makes Ireland appear a highly candidate-focussed system. Looking next down the rightmost column, however, we get a slightly different picture. About 46 per cent said they would support the same candidate running for a different party – these certainly do seem to be people who put candidate over party. About 37 per cent said they would not, while 17 per cent said that this would depend on which other party was involved.
[table 7.1 about here]
Looking inside table 7.1 we first see a peculiar group of people: the 6 per cent of apparently bewildered respondents who said that party is more important than candidate when deciding how to cast their first preference vote, but who also say that they would definitely give their first preference vote to the same candidate if he or she ran for a different party.
Apart from this, we see that about 40 per cent of voters can be regarded as really committed supporters of candidates rather than parties. The candidate is more important for them than the party, and they would support the same candidate if he or she ran for any other party. Of the remaining 22 per cent of voters who said that candidates were more important than parties when allocating their first preference votes, about half said they would not support the same candidate running for any other party – these people seem much more like party than candidate supporters, choosing between different candidates of the party they want to support. About half said this would depend upon the party concerned – these voters clearly have parties in mind when allocating their first preference vote, even if the candidate is the most important factor.
The really hard-core party supporters make up about 27 per cent of respondents – these are the people who choose parties rather than candidates, and would not support the same candidate running for any other party. Another 6 per cent or so of respondents are party voters, but might support the same candidate running for another party – depending on what party that was.
In short, on the matter of whether voters choose candidates or parties, the INES allows us to carve up the electorate in 2002 as follows. About 40 per cent were dyed in the wool supporters of some particular candidate, and would have followed that candidate across party lines. About 37 per cent of voters were party supporters, giving a first preference to a candidate only if he or she runs for that party. Another 17 per cent or so of voters take both party and candidate into account, being willing to consider supporting the same candidate running for a different party, but conditioning this on which party is involved. The remaining 6 per cent show contradictory views of the kind we often encounter in surveys. Crudely, about 40 per cent seem to be candidate voters, about 40 per cent party voters, with the remaining 20 per cent balancing to two motivations. Interestingly, age does seem to have a systematic impact on voting motivation, with older people more likely to be committed party voters. About 30 per cent of 18-25 year olds said they would not support the same candidate running for another party, as opposed to 50 per cent of those aged over 65.
Determinants of party support
As we saw in the introduction to this chapter, the Irish party system has been has been characterised as having no ‘social bases’ – a product largely of the cross-class appeal of the largest party, Fianna Fáil, and the long-term weakness of the Labour Party. The INES allows us to expand considerably on these conclusions, asking more people more detailed questions on their voting habits than ever before. This new information allows us for the first time to begin at the beginning, with the political socialisation of voters, when they were children, in the family home, before going on the explore the impact of a range of different social characteristics on Irish voting behaviour.