Mobilising the 'Women's Vote': The Gender-Generation Gap in Voting Behaviour

Transcription

1 Mobilising the ''s Vote': The Gender-Generation Gap in Voting Behaviour BY PIPPA NORRIS THIS article explores the nature of gender differences in British voting behaviour. The first part outlines trends in British general elections from 9 to 992. Rather than a simple 'gender-gap', we argue that it is more useful to talk about a 'gender-generation gap'. Instead of acting as a homogeneous voting bloc, younger and older women are deeply divided in party preferences. Older women remain more Conservative than older men, with significant consequences for party fortunes given the number of older women in the electorate. Nevertheless, younger women are more left-wing than their male counterparts. Subsequently, we consider whether this pattern is found in other countries in recent years, using data from the 99 EuroBarometer. This suggests considerable variation cross-nationally in the gender-gap, with women more right-wing than men in Britain, Australia, Luxemburg and Italy while women are more left-wing in Germany, Spain, Portugal and the United States. The comparison confirms that gender differences vary by generation across Europe. Lastly, we consider the implications of these developments for patterns of party competition in pursuit of 'the women's vote' and the consequences for the future of the Labour and Conservative Parties. Before we can examine trends, we need to clarify our terms. The term 'gender-gap' is a catch-all phrase which can be used to refer to widely different phenomena, including divergences between women and men in turnout, party identification, political attitudes, social values and policy concerns, at mass and elite levels. The focus here is the most common usage, namely difference between women and men in their voting choice, whether in a left-wing or right-wing direction. For this analysis the gender-gap will be summarised as the differences between the two-party (Conservative minus Labour) vote lead among women and men: Gender-gap = (% Con - % Lab) minus (% Con - % Lab). By this measure a positive gender-gap signifies women are more rightwing than men, while a negative gender-gap indicates they are more left-wing. The two-party vote lead is the most useful and convenient summary measure since in the postwar period there have been only minor gender differences in support for third parties. Differences in the social and economic situations of men and women, Oxford University Press Downloaded from

2 Parliamentary Affairs Conservative Vote 992 Labour Vote as well as their attitudes and values, will be treated here as independent variables which are possible explanations for their party choice. The older literature often followed a reductionist path, tending to explain away gender differences as the product of religion, longevity or socioeconomic status. Instead, it can be argued, we need to test for direct and indirect effects. That is, gender may have a direct effect on voting behaviour. For example, if women and men differ strongly in the priority they give to environmental issues, this may influence their support for the Green Party. Alternatively, the life-styles of men and women differ substantially in many regards occupational class, personal income, family responsibilities, union membership, and religiosity. Through these structural differences gender may have an indirect influence on voting choice. Based primarily on material derived from the series of British Election Studies from 9 to 992, and Euro- Barometers in 9 and 99, we can consider some of the alternative factors linking gender to voting behaviour. Trends in the gender-gap First we need to establish the main trends over time in Britain. In 9, when the female franchise was first granted, it was widely expected that women would favour the Conservatives, given the party's traditional emphasis on Church and family, while more men were expected to support Labour due to the party's links with trade unions in traditional heavy industries like mining, steel manufacturing and the shipyards. In the absence of systematic evidence, the prewar situation has to remain speculative. Nevertheless, postwar trends can be analysed from the series of Gallup polls, which started on a national basis in 9, 2 followed by the more detailed series of British Election surveys. The evidence for confirms the conventional wisdom. If we compare women and men voters in the Gallup series, we find the Conservatives consistently did better among women, while Labour proved stronger among men (see Figure and Table ). Indeed, if Britain had continued with an all-male franchise, all other things being equal, there would have been an unbroken period of Labour govern- Downloaded from

3 Gender-Generation Gap in Voting Behaviour Vote by Gender 9-92 Coo Lab lib Gender Gap Note: The gender-gap is calculated as the difference in die Con-Lab lead for women and men. A positive gender-gap indicates that women are more Conservative than men. Source: Gallup Polls, 9-59; BES ments from 9 to 979. The Conservative lead among women was significant throughout the 950s, when the gender-gap ranged from to 7 points. In 9 Wilson entered No.0 in part due to a stronger pro-labour swing among women than men. In 970 men favoured Labour, while women gave a greater lead to the Conservatives, producing a gender-gap of points and victory for Heath. As a result, the political science literature in the 90s and 970s focused on explaining why women were more conservative, noting that the pattern in Britain was also evident in early voting studies in the United States, Italy, France and West Germany. Indeed, this became the accepted view in political science, as Randall notes: 'Up to the 970s, women were apparently more inclined than men to vote for conservative parties in every country for which information is available including not only Greece, Belgium, Switzerland and the Netherlands, but also Sweden and Finland.' 5 During the 90s the Conservative edge among women became statistically insignificant, with a gender-gap of, 2 and percentage points, in respective elections from 979 to 97. As Rose and McAllister summarised the position following the 97 general election: 'Notwithstanding traditional theories of women favouring the Conservatives and feminist theories indicating that women ought to vote Labour, gender has no influence upon voting in Britain today. The reason is straightforward: on matters that are salient to voting, men and women tend to share similar political values. On most major political issues men and women divide similarly along lines of party or class, not gender.' In the light of these trends, the return of a small but significant gender-gap in the 992 British general election, with women more conservative than men, was unexpected. The final campaign polls by Harris, NOP and MORI confirmed this pattern (see Table 2). These found the Conservatives enjoyed a slight edge among women voters, Downloaded from

4 Parliamentary Affairs Harris Gender-gap MORI Gender-gap NOP Gender-gap BES Gender-gap 2. Estimate! of tbe gender-gap in 992 Con 2 2 Lab LD Note: Not all die above figures sum due to correction of rounding error. Con/Lab Lead although estimates of the size of the gap varied from to points. In the 992 British Election Survey the Conservative lead over Labour was.% among all men but.% among all women producing a gender-gap of points. One of the most striking aspects of this phenomenon is the fact that the gender-gap reverses by generation (see Table ). In the 992 general election, younger women (those aged under 0) gave a lead to Labour, while younger men shifted sharply towards the Conservatives, producing a - point gender-gap. If we break this down further, the gap is 0 points amongst the -2 age group, with young men voting far more Conservative (5%) than young women (%). Among the middleaged the pattern reverses, with women more Conservative than men. In the same survey the Conservative edge among older voters created a substantial gender-gap of points. This is not a new development. Figure 2 shows trends among the youngest and oldest groups in 992. This reflects patterns from 9 where younger women have proved the least Conservative in successive elections, except for the period from 970 to October Older women have usually been the most Conservative, except for the 97 election. The consistency of this trend indicates that there is a well- All Young (>0) Middle (0-) Older (5) 5 Con Vote by Age Group Lab LD Gender Gap Coeff.of As toe.05*.2".07*.0* Note: Gamma is used to test the coefficient of association."=sig.0 *=rig.05. A positive gender-gap indicate* that women are more Conservative than men, a negative gap that women are less Conservative. Downloaded from

5 Gender-Generation Gap in Voting Behaviour 7-2 Conservative Voting Groups 992 million votes Labour Voting Groups 992 H M ^ i H million votes i established 'gender-generation' gap. If so, this pattern should not be attributed to short-term factors specific to the 992 campaign, such as the influence of particular issues, party policies or political leaders. Instead we need to consider long-term explanations. This pattern has a significant impact on party fortunes. The overall gender-gap among all women was small in percentage terms, compared with class or housing cleavages, but it was statistically significant, it was replicated in all the final polls, and it was sizable in delivering ballot box votes since women are the majority of the population, with slightly higher rates of voter participation. There were. million older women voters in the last election, compared with 2. million men in the equivalent age group (over 5). The Conservatives received million votes in the 992 general election. Of these.9 million were from older women (see Figure ). No party can afford to discount so large a group of voters. Given the narrowness of the government's parliamentary majority, it could be claimed that Mr Major was returned to No.0 in part due to the older 'woman's vote'. Cross-national trends Turning to cross-national trends, the conventional wisdom was established by Maurice Duverger in The Political Role of (955), Conservative Vote by Gender »_ ^^<S^<^S. r Downloaded from

6 Parliamentary Affairs which found that women voters were slightly more right-wing than men in many European countries, including Italy, Norway, France, and Germany. Vicky Randall suggests that during the 90s in most countries women voters were consistently more conservative than men by a small margin, according to the available evidence. Female conservatism was commonly explained by their greater longevity and religiosity, and their lower trade union membership, rather than by gender as such. 9 Yet in the last decade women in the United States have become more left-wing than men, in a well-established realignment with profound consequences for gender politics. 0 Similar patterns have been suggested for Scandinavia, although others report no consistent gender-gap in this region. A thorough review of the evidence in Nordic politics concludes that gender differences in party choice are modest in Denmark, Norway, Sweden and Finland, and only significant in Iceland where the 's Alliance party is heavily female-dominated among its supporters. Nevertheless, the study found that differences in party choice were larger among younger than older voters, which may lead to an emerging gender-gap in future. 2 Others found that during the 90s gender differences in electoral behaviour converged in Germany, Italy, and France, with women and men voting increasingly alike. We can examine recent trends in the gender-gap by analysing the 99 EuroBarometer. The results are not easy to interpret on a consistent basis in different European countries for two reasons. The measurement of any gender-gap depends heavily on how particular parties are classified along the left-right scale. With some parties this may be obvious. With others, such as the Green parties, regional parties and European Liberal parties, this is more difficult. The classification used here is based on party affiliation to European parliamentary groups. Moreover, we would expect any gender differences in voting choice to be relatively modest. The gender-gap, even in the United States, is not large in terms of the percentage difference between women and men. But it is politically significant since women are over half the electorate, and even a modest ripple in their voting choice can produce large shifts in levels of party support. Bearing in mind these qualifications, the 99 EuroBarometer provides an indication of women and men's voting choice. This can be compared against published data from the 992 US presidential election and the 99 Canadian and Australian general elections. The analysis for the major parties shows striking cross-national variations (see Table ). were more left-wing than men in Canada, West and East Germany, Portugal, Spain and Denmark, as well as the United States. By contrast, in Britain, Australia, Luxemburg, Italy, Ireland and France they proved more right-wing than men in their voting choice. In the remainder of countries there were no significant gender differences. The summary gender-gap varied substantially in strength as well as direction, Downloaded from

7 Gender-Generation Gap in Voting Behaviour 9 Britain Australia (a) Luxemburg Italy Ireland France Greece Belgium Netherlands Denmark W. Germany Spain United States (b) Portugal E. Germany Canada (c) Vote for the M»jor Parties of Right and Left Left-wing Vote Gap Right-wing Vote Gap Left-Right Lead Sumnu Genderl A All Note: Percentage vote for ma)or parties in rhe 99 European electioru, the 99 Australian and Canadian general elections and the 992 US presidential election. A negative gender-gap indicates that women are more right-wing. A positive gender-gap indicates that they are more left-wing. with Britain proving the most conservative of all the countries under comparison. Given the problems of classifying parties, we need to see whether this pattern is replicated using alternative measures. In the EuroBarometer data we can compare where voters placed themselves on a ten-point left-right scale, where a low score represents most left-wing, as a simple measure of their ideological identity. In the 9 EuroBarometer, gender differences in this scale were modest, with 5. as the mean position for women compared with 5.2 for men. Yet there was evidence of a gender-generation gap. Younger women positioned themselves slightly left of men, while older women positioned themselves slightly right of men. A decade later, this gender-generation gap was even more marked (see Table 5). In 99, younger women were slightly more left-wing than younger men. This pattern reverses among the middle-aged, while older women and men were very similar in their self-placement. Turning to the national comparison, in Germany, Denmark, France and the Netherlands women were more left-wing than men. By contrast, in Portugal, Ireland, Italy and Britain women were more right-wing than men. In the remainder there was no difference between the sexes. By 99, across the whole of Europe, overall women saw themselves as slightly more left-wing than men, although we can conclude that any gender-gap did not point in a consistent direction across all member states. We can conclude from these results that there are significant crossnational differences in the gender-gap, both in terms of left-right ideology and in voting choice, which need to be explored further. The Downloaded from

8 0 Parliamentary Affairs All W. Germany Denmark E. Germany France Netherlands Greece Spain Luxemburg Belgium N. Ireland Britain Italy Ireland Portugal Younger Middle-aged Older 5. Left-Right Ideological Self-Placement 5-25 S2S Note: The scale varies from left () to right (0). 5J Gap results suggest that the conventional wisdom about women's greater conservatism across Europe is now no longer valid and European parties of the centre-right can no longer rely upon the automatic loyalty of women voters. Nevertheless, we remain a long way from explaining the gender differences which are evident. The consequences for party competition The development of the gender-gap in the United States has had a major impact on patterns of party competition in America, heightening the salience of gender-related issues like affirmative action, abortion and sexual harassment. The organised feminist movement, and women politicians, have seized on the gender-gap to advance their agenda. It may also be part of a backlash. The 99 elections were framed through the prism of gender, with endless stories in the media about the 'angry white male' who had deserted the Democrats, despite good evidence that the gender-gap in the 99 elections was the same one which has persisted since the early 90s. The gender-gap in most presidential and congressional elections has never been great in the region of to 0 percentage points but it has been politically significant. The reasons are that the women's movement mobilised around this development; it affects millions of votes; these votes are dispersed across every electoral district; the gender-gap cannot be explained (and therefore modified) by a single issue; and it is a relatively recent phenomenon. As a result, the Democrats and Republicans have sought to compete hard for women's ballot box power, with only a hazy understanding of what steps are necessary to retain or regain their votes. Like the cleavages of region and race, gender has now become an established part of conventional candidate strategy. Downloaded from

9 Gender-Generation Gap in Voting Behaviour In Britain, during the 992 general election, the major parties also attempted to win women voters, but their efforts were a pale shadow of party competition in the United States. The Conservatives argued equality of opportunity was preferable to positive discrimination and stressed the government's record on equal rights in employment, education and training, plus private sector initiatives such as Opportunity 00. The Liberal Democrat manifesto proposed to fight sexual discrimination through a Bill of Rights and to bring more women into political office through the introduction of proportional representation, although gender equality was not treated as an issue of high priority. The greatest effort was made by Labour. The party's Shadow Communications Agency had recognised the need to target women after the 97 general election. It was felt that Labour suffered from a predominately male image, particularly at party conferences, partly due to its strong links with the trade unions. Given the shrinking workingclass and trade union base, women were identified as one of the target groups where Labour could expand its support. 's organizations within the party took advantage of this opportunity to advance their own agenda. Accordingly, Labour's strategy was to craft publicity material like the manifesto and the conference backdrop to present a socially diverse picture; to develop policies to strengthen women's rights, such as the proposed Ministry for ; and to increase the visibility of women at all levels of the party, including record numbers of women candidates for Parliament. 7 Nevertheless, the impact of this strategy was modest. The number of women candidates increased across all parties. Labour's policy proposals were tucked away in the manifesto and rarely surfaced in the campaign. There was probably more media coverage of the leader's wives than of policies towards child-care or equal pay. During the campaign Labour chose traditional male party spokesmen to present its policies. One study of television coverage on the main evening news counted the number of broadcasts during the campaign. Out of Labour appearances, only three were by women (Harriet Harman and Margaret Beckett). In contrast, in the middle of the campaign Mrs Thatcher played a prominent role aimed at restoring grassroots Conservative morale. As noted in the official Labour Party inquest into the election by the general secretary Larry Whitty, Labour's efforts proved insufficient to attract more women voters into its fold, particularly among the middle- aged and older. 9 In the run up to the next general election, the Labour Party is again trying to target women's votes. As discussed in other articles in an earlier issue, this is one factor behind the adoption of all-women short-lists in target Labour seats. Other policies include the pledge of a Ministry for to prioritize women's issues. Weakening the links with traditionally male-dominated trade unions, and the attempt to create a more participatory party at the grassroots, along with the move towards the centre ground and a Downloaded from

10 2 Parliamentary Affairs stronger appeal to traditional 'family values', are other initiatives which may appeal to women. It remains to be seen whether Labour's strategy proves the key which will switch the traditional gender-gap in the next general election. For an earlier version of this argument, tee P. Norm, The Gender Generation Gap', British Elections and Parties Yearbook 99 (Harvester Wheatsheaf, 99). 2 H. Durant 'Voting Behaviour in Britain 9-' in R. Rose (ed), Studies m British Politics (Macmillan, 2e 99). See I. Crewe, N. Day and A. Fox, The British Electorate 9-7 (Cambridge University Press, 99). See M. Duverger, The Political Role of (UNESCO, 955); D. Buder and D. Stokes, Political Change in Bruam (Macmillan, 97); J. Blondel, Voters, Parties and Leaders (Penguin, 97); R. Rose 'Britain: Simple Abstractions and Complex Realities' m Electoral Behaviour (Free Press, 97); P. Puber, Political Representation and Elections in Britain (Allen and Unwin, 97). 5 V. Randall, and Politics (Macmillan, 2e 97). R. Rose and I. McAllister, The Loyalties of Voters (Sage, 990), p See S. Baxter and M. Lansing, and Politics: The Invisible Majority (University of Michigan Press, 90), p. 57. For me 9 election see P. Findlay, S. Griffiths and M. Franklin Generation and Gender m British Voting (Strathclyde Papers on Government and Politics, 95). Randall, op. cit 9 Baxter and Lansing, op en. 0 For reviews of die American literature see C. Mueller, The Politics of the Gender Gap (Sage, 9) and The Gender Gap and 's Political Influence' Annals of the American Academy of Political and Social Science, May 99. L. Karvonen and P. Selle, in Nordic Politics (Dartmouth, 995); E. Haalvio-Manilla et al Unfinished Democracy (Pergamon, 95). 2 M. Oskarson, Karvonen and Selle, op. cit. See A. Appleton and A. Mazur, Trance' (Diagram 5.), and E. Kolinsky, 'Germany' (Table.), in J. Lovenduski and P. Norris, Gender and Party Politics (Sage, 99). P. Norris, The Gender Gap: A Cross National Trend?' m C. Mueller, The Politics of the Gender Gap, (op. cit). 5 To consider explanations see P. Norns in M. Simms, The Future of the Australian Party System (Allen k Unwin, Australia, 99). P. Norris, 'Introduction:, the Media and Politics' and E. C Ladd, 'Media Framing of Election Results' in P. Norris (ed),, the Media and Politics (Oxford University Press, 99). 7 See V. Atkinson and J. Spear, The Labour Party and : Policies and Practices', M. Smith and J. Spear (eds), The Changing Labour Party (Routledge, 992). See P. Golding, Economy and Polls Hogged Coverage', The Guardian, See P. Webster, "Labour Says Failure to Win Voter's Trust Lost Election', The Tones, Downloaded from