Portal de Periódicos da CAPES

A Conservative Revolution: The Electoral Response to Economic Crisis in Ireland

2014; Taylor & Francis; Volume: 24; Issue: 2 Linguagem: Inglês

10.1080/17457289.2014.887719

1745-7297

Michael Marsh, Slava Mikhaylov,

Populism, Right-Wing Movements

AbstractThe 2011 election in Ireland was one of the most dramatic elections in European post-war history in terms of net electoral volatility. In some respects the election overturned the traditional party system. Yet it was a conservative revolution, one in which the main players remained the same, and the switch in the major government party was merely one in which one centre-right party replaced another. Comparing voting behaviour over the last three elections we show that the 2011 election looks much like that of 2002 and 2007. The crisis did not result in the redefinition of the electoral landscape. While we find clear evidence of economic voting at the 2011 election, issue voting remained weak. We believe that this is due to the fact that parties have not offered clear policy alternatives to the electorate in the recent past and did not do so in 2011. Notes1. This is the combined vote of the centrist left Labour (19.4%), the nationalist left Sinn Fein (SF) (9.9%) and the United Left Alliance (2.7%). If we add the Green Party (1.8%) it would rise to 32.8%. In 2007 these combined parties won only 21.6%, with the Greens contributing 4.6%.2. This was a post-election study of 1,853 respondents with interviewing completed a few weeks after the election. Households were selected using a random route method starting from 320 locations drawn from all 43 constituencies. Interviewees were chosen within households to fit demographic quotas.3. We included party identification to allow for the possibility that many voters would view the world through partisan spectacles (Campbell et al., Citation1960; Duch et al., Citation2000; see also Evans & Andersen, Citation2006, for a stronger critique). For a contrary view see Green et al. (Citation2002). Partisanship – measured by the question asking are you close to any political party – is very low, with only about a quarter of voters admitting to such an identification, and dropping to 20% in 2011, as FF partisans vanished. Clearly, this measure cannot be seen as wholly independent of recent experiences. We opted for a simple model without interactions (c.f. Marsh & Tilley, Citation2010) because including the responsibility measure and the assessment of an alternative ran the risk of overcorrecting for a partisan bias.As we see, the partisanship variables seem to be doing most of the work. Without them the pseudo R2 is much lower, 0.08, 0.04 and 0.05, respectively.4. The parameters of multinomial models are not directly interpretable but their significance can be immediately discerned. Much of our discussion below specifically revolves around the analysis of significance of individual coefficients (and overall model performance). However, in multinomial models coefficients can be calculated only in relation to the baseline category (Fine Gael in our case), while we would prefer to have the effects for all parties. This can be achieved by changing the focus from the logit coefficients to the marginal effects that can be calculated for all categories (parties). The marginal effects in multinomial models capture the impact of a change in the predictor on the probability of observing each of several alternative outcomes. The table presents average marginal effects (AMEs) for each of the alternatives with corresponding uncertainty estimates. The reported results are derivatives (or discrete first-differences for factor variables) of the response with respect to the key predictors, i.e. it captures a change in the response for a change in the predictor.5. All other parties are included with Independents as Others. The numbers voting for each of these are too small to analyse usefully in this manner. Independents are a very disparate group and we have not shown the independents/others effects, but they are generally not significant.6. Full results from this estimation are not shown here to preserve space. They are available from the authors upon request.7. A special feature of 2011 that attracted comment was the degree of anger felt by many voters as they were let down by the government and the financial elite, and politicians in general, with large salaries and generous expenses and very comfortable pensions and severance payments. On a scale of 1–5, where 5 is “extremely angry” the median voter was at 4: very angry at “how things were going in the country these days”. Including this emotion adds only marginally to the fit of the models run above, but does suggest SF and Independent voters were even angrier than those of FG and Labour.8. The results are available from the authors upon request.9. I would like you to look at the scale from 0 to 10 below. A “0” means government should CUT TAXES A LOT and SPEND MUCH LESS on health and social services, and “10” means government should INCREASE TAXES A LOT and SPEND MUCH MORE on health and social services. Where would you place yourself in terms of this scale? This scale has been reversed in the discussion here, so that “left” is 0 and “right” is at 10.In 2011, arguably, the choice was not so much one between increasing tax or cutting spending, but how much tax should increase and how far spending should be cut. For this reason the INES asked a second question: I would like you to look at the scale from 0 to 10 on this card. A “0” means government should MAINTAIN TAXES and SPEND LESS on health and social services, and “10” means government should INCREASE TAXES A LOT and SPEND THE SAME on health and social services. Where would you place yourself in terms of this scale?There is very little difference in either the distribution, or the effects on vote between these two wordings.10. We find much the same picture if we look at another measure designed to tap positions on a left–right scale. There was an increase in support for more rather than less regulation in 2011, for instance, although there was also an increase for private rather than public enterprise.11. Full results are available from authors upon request.12. Full results are available from authors upon request.13. A similar analysis was carried out on the 1990 World Values Study, but with different measures and results that suggested bigger differences between parties than we find here: see Hardiman and Whelan (Citation1994).14. This was the final wave of a five-wave study, and even though it was supplemented to achieve a more representative sample, it is likely that the voters in 2007 were more sophisticated than those in 2002 or 2011.15. Such self-reporting may be considered unreliable, and may reflect a perception that parties should be more important rather than real motivation. However, there is separate evidence that more voters did give a priority to parties. Those voting a “straight ticket” – giving a preference to all candidates for one party before those of a second party – did rise in 2011. This comes from the mock ballots filled in by election study respondents. (For a discussion of this method see Marsh et al., Citation2008.) Moreover, those who said they were more party-centred did fill in their ballots in a more party-centred way. Using the combined measure above, only 34% of candidate-centred voters cast a straight ticket compared with 55% of party-centred voters. Comparable percentages for 2002 and 2007 are 23 and 58, and 27 and 58.16. For 2011, Pseudo R2 is 0.028 and 0.019 for party- and candidate-centred voters respectively using the variables in Table 4, column 3 without party attachment. If a behavioural measure (and the number of cases is necessarily smaller here) it is used, R2 for those voting a straight party ticket is 0.043 and 0.025 for those who do not.17. Lowe et al. (Citation2011) define the “right” side of the dimension state involvement in economy as a combination of positive reference to free enterprise, economic incentives, economic orthodoxy, welfare state limitations, and negative references to protectionism. The “left” side is defined by positive references to market regulation, economic planning, protectionism, controlled economy, nationalization, welfare state expansion, education expansion, and labour groups. Position on the dimension is then scaled as the empirical logit transform of the difference between right and left. Thus more negative positions reflect more “leftist” positions.18. Lowe et al. (Citation2011) define the “right” side of the dimension state-provided services as a combination of positive reference to welfare state limitation, education provision limitation. The “left” side is defined by positive references to welfare state expansion, and education provision expansion. Position on the dimension is then scaled as the empirical logit transform of the difference between right and left. Thus more positive positions reflect more “rightist” positions.19. General left–right is a combination of 26 categories from manifesto analysis, capturing a variety of left and right issues.