Brexit and prosperity but defeat: the economic vote conundrum in the 2020 Irish election

ABSTRACT

At face, Ireland's economy had staged a remarkable recovery by 2020 since the devastating impact of the Global Financial Crisis. The economy was the fastest growing in Europe, unemployment had reached record lows, and Ireland's debt was back to its lowest level since 2009. The traditional economic voting model assumes voters punish outgoing governments for poor economic performance by voting against them but rewards incumbents for a sound economy by voting for them. Nevertheless, Irish voters delivered a stunning rebuke to the Fine Gael government in 2020, registering an incumbent administration's sixth-worst performance since 1932, raising questions about the applicability of the economic vote. Using the 2020 Irish National Election Study, our contribution unpacks this apparent puzzle. We uncover that macroeconomic conditions were less rosy than at first sight, a pattern recognised by voters. We find the economic vote was alive and well, with voter economic perceptions, and their views on income redistribution (and taxes/spending) having a potent effect on the vote. Brexit, as an economic issue, however, was not influential in shaping the vote. The Fine Gael challenge, and the answer to the conundrum, was most voters perceived the economy was lackluster and the government was on the wrong side of economic policy preferences of most voters.

KEYWORDS:

• Economic voting
• positional economics
• Brexit
• valence economics

Acknowledgements

We presented an earlier version of the paper at the 2020 PSAI Annual Conference (online), and we appreciate the helpful comments of attendees. We thank David Farrell and Johan A. Elkink for developing the 2020 INES, sharing the data with us, and their valuable critique. We also acknowledge the valuable feedback from Deirdre Tinney and two anonymous reviewers. As ever, all remaining errors are our own.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Another debate rages about whether voters are retrospective or prospective. We do not address this issue here but accept the conventional wisdom in the valence paradigm that voters are principally retrospective.

2 It refers to inferring the behaviour of the electorate from its parts (the voters) – see Dassonneville & Lewis-Beck, 2014.

3 For robustness, we applied a separate political weight variable based solely on reported vote, and thus including larger Ns. We detail these analyses in Table C3, Appendix C (valence model) and Table C4, Appendix C (traditional positional model). We detect no deviations for the variables of interest declared in the text.

4 For robustness, we specify a composite model incorporating valence and positional measures in the one model with the same dependent variable operationalisation used for the positional economic models (see Table C5, Appendix C). We discover the traditional positional economic measure retains statistical significance (p < 0.05) for voting for both rightists in opposition and government. The sociotropic valence measure also retains statistical significance for the vote for the government, sustaining our finding of a multidimensional economic vote.

5 Sceptics might argue some of the parties constituted as unconstrained right are to the right of Fine Gael on the ideological spectrum. For robustness, we estimate positional models where the Irish Freedom Party, Irish National Party, and Renua are classified as to the right of Fine Gael (see Table C6, Appendix C). We detect no deviations for the variables of interest declared in the text.

6 For robustness, we tested the positional economic voting thesis using a measure of voter preferences between tax cuts and more spending on social services. The results are almost identical, giving confidence that a positional economic dynamic is at play– see Table C10, Appendix C.

7 For robustness and due to the potential for Brexit to also act as a valence issue, we specify the Brexit variable in a valence model (see Table C7, Appendix C). We detect no deviations for the variables of interest declared in the text.

8 For robustness, we specify our models with alternative specifications, including capturing respondent left-right ideological predilection instead of partisanship, and including respondent left-right ideological predilection and partisanship simultaneously. We detect no deviations for the variables of interest declared in the text (see Table C2, Appendix C for valence model; and Tables C8-9, Appendix C for traditional positional model).

9 A similar pattern is seen in the data from the exit poll. When asked whether or not the statement ‘I have benefitted from the improvement in the economy in recent years’ applies to them or not, only one third of voters reported it did, while two thirds said it did not apply to them – see Table A3, Appendix A.

10 For robustness, we explored which respondents were more likely to provide positive sociotropic and egotropic assessments (see Tables C1, Appendix C).

Additional information

Notes on contributors

Michael S. Lewis-Beck

Michael S. Lewis-Beck is F. Wendell Miller Distinguished Professor of Political Science at the University of Iowa. His interests are comparative electoral behaviour, election forecasting, political economy, and quantitative methodology. His research has appeared in journals such as The American Journal of Political Science, Political Behavior, The American Political Science Review, Journal of Politics, and Comparative Political Studies. He has authored or co-authored over 285 publications, including the books Economics and Elections, The American Voter Revisited, The Danish Voter: Democratic Ideals and Challenges, and Applied Regression.

Stephen Quinlan

Stephen Quinlan is Senior Researcher at the GESIS – Leibniz Institute for the Social Sciences, Mannheim, Germany, and is Project Manager of the Comparative Study of Electoral Systems (CSES) project. His research interests are comparative electoral behaviour, social media’s impact on politics, and election forecasting. His research has appeared in journals such as Electoral Studies, Information, Communication, and Society, Party Politics, European Political Science Review, and the Journal of Elections, Public Opinion, and Parties.