Portal de Periódicos da CAPES

The Collapse: Economic Considerations in Vote Choice in Iceland

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

10.1080/17457289.2014.889699

1745-7297

Indridi Indridason,

Fiscal Policies and Political Economy

AbstractThis paper examines the consequences of the collapse of the Icelandic banking system in the fall of 2008 on voters' support for the political parties. The literature on economic voting has demonstrated that voters hold governments accountable for past economic outcomes, and such concerns should be especially salient in times of large-scale economic crisis such as that experienced by Iceland. In such cases, where the cause is more likely to be seen as the consequence of policies having to do with organization of the economy and, in particular the banking sector, rather than a consequence of short term economic management, the question of who voters hold accountable is of particular interest. Notes1. In combination the three banks constituted about 85% of Iceland's banking system (Gylfason, Citation2012).2. The parties represented in Althingi were: The Independence Party (conservative), the Progressive Party (center-right, former farmer's party), The Social Democratic Alliance (center-left), the Left Movement – Greens, and the Liberal Party (center-right).3. Alternatively, greater concern with unemployment may cause voters to lean left (Kwon, Citation2008, Citation2010).4. The data on expected household income, consumer confidence, and support for government come from Capacent's monthly surveys and are available from http://datamarket.comDatamarket.com.5. We define government coalitions solely in terms of their party composition. The first coalition according to our definition, the coalition of the Independence Party and the Progressive Party, would normally to be considered as a series of coalitions. The coalition was first led by Davíd Oddsson (IP) who handed the reins of power to Halldór Ásgrímsson (PP) in September 2004 as a part of the parties' agreement to renew their agreement to govern together after the 2003 parliamentary election. Finally, Geir H. Haarde (IP) took over as the Prime Minister on June 15, 2006.6. The coalition of the Social Democrats and the Left-Green Movement was renewed after a parliamentary election on April 25, 2009, now as a majority coalition, on May 10, 2009.7. The results of tests for non-stationarity depended on whether the whole time-series was considered or if it was broken down by government. The first two models use Government Support as the dependent variable and include lags of Government Support to account for autocorrelation. In the next four columns the data are first differenced, i.e., Δ Government Support is modeled as a function of change in the independent variables.8. The questions address current economic and labor market conditions and expectations about economic conditions, labor market conditions, and household income in six months. It bears noting that measures of consumer confidence may be endogenous to the political context (De Boef & Kellstedt, Citation2004).9. The effect is statistically significant if the squared value of Expected Income is also included in the model.10. As many, for example Evans & Andersen (Citation2006), Johnston et al. (Citation2005), Wilcox & Wlezien (Citation1993), and Wlezien et al. (Citation1997), have shown there are reasons to believe that partisanship colors evaluations of government performance rather than the other way around. The data used here are evaluations of government performance but the same concerns apply and, thus, the interpretation of the results regarding government performance should be taken with a grain of salt.11. The models for different years are not fully comparable as some questions were not included in all the surveys. That is obviously the case for questions related to the economic crisis but that was also true about some of the other variables, e.g., Best Party for Economy.12. The effect is calculated by varying the value of each independent variable from its mean standard deviation to mean standard deviation while holding other variables fixed at their means except for dichotomous variables, which are held fixed at their mode. The effects of the party-specific variable are not summarized here in a similar manner, as the values of these variables are not independent across parties, i.e., only one party can be considered the best party for the economy. Thus, the effect of being considered the best party for the economy for, say, the SDA depends on which party is initially assumed to be the best party for the economy.13. Replacing the variable with one based on a question about managing economic recovery yields substantively similar results.14. The exception is that the effect of IP-SDA Performance was positive for the PP (+6.8% pts) and the respondents were less likely to vote for the SDA (−10.0% pts). It is possible that the respondents that rated the performance of the IP–SDA coalition more highly were also more likely to consider the responsibility for the crisis to lie with non-political actors, i.e., the banks and the FSA, and were less likely to hold the political parties accountable for the crisis. These voters may, therefore, also be less likely to hold the Progressive Party accountable, despite the fact that it had been in government from 1999 to 2007, and support it.15. The average rating for the IP–PP government in 2007 was 1.69 whereas its successor, the IP–SDA cabinet, had an average rating of 0.79.16. Following the 2013 election, the government of the IP and the PP halted the accession negotiations.17. Note that the effects are estimated by holding the values of responsibility at their mean for the other parties but typically holding one party more responsible means holding some other party less responsible. Thus, the estimated effect is likely underestimated.18. In 2009 the number of respondents that reported a vote for the LP or the CM was relatively small while many respondents had a hard time placing those parties on the left–right policy dimension. The results are very similar when the analysis is restricted to the four main parties.