Special elections in alternative vote electoral systems: Exploring turnout and the vote in Irish by-elections 1923–2019
| Author | Stephen Quinlan |
| DOI | http://doi.org/10.1177/13691481211045730 |
| Published date | 01 February 2023 |
| Date | 01 February 2023 |
| Subject Matter | Original Articles |
https://doi.org/10.1177/13691481211045730
The British Journal of Politics and
International Relations
2023, Vol. 25(1) 156 –177
© The Author(s) 2021
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/13691481211045730
journals.sagepub.com/home/bpi
Special elections in alternative
vote electoral systems:
Exploring turnout and the vote
in Irish by-elections 1923–2019
Stephen Quinlan
Abstract
Most literature on special elections has focused on first-past-the-post contests and on the
performance of governments. Turnout, candidates, and how the electoral system impacts the
result have received less attention. This contribution fills these voids by exploring special elections
in Ireland, elections conducted under the alternative vote system. Taking a multifaceted approach,
it investigates the correlates of turnout, the impact of candidates and the decisive effect of lower
preferences, while also testing multiple explanations of government performance. I find Irish special
elections live up to the by-election truisms of lower turnout and government loss. Government
performance is associated with national economic conditions. By-election victory is more likely
among candidates with familial lineage and former members of parliament. Where they come into
play, one in five candidates owe their victory to lower preferences.
Keywords
alternative vote, by-elections, government performance, Ireland, special elections, turnout
Introduction
Special elections, by-elections, by-polls or Elezioni Suppletive – these contests, held
when a seat becomes open during a parliamentary term, are routine in many polities.1
Feigart and Norris (1990: 185) found over 60 states employ them, with Italy recently
introducing them under its 2017 electoral reform. Despite their prevalence, special elec-
tions have received less attention than other so-called less important elections like
European parliament elections (Marsh, 1998; Okolikj and Quinlan, 2016; Reif and
Schmitt, 1980; Schmitt, 2005), midterm polls (Campbell, 1985, 1986; Erikson, 1988;
Tufte, 1975) or local and regional contests (Carsey and Wright, 1998; Chubb, 1988;
Marien et al., 2015).
Department of Monitoring Society and Social Change, GESIS Leibniz Institute for the Social Sciences,
Mannheim, Germany
Corresponding author:
Stephen Quinlan, Department of Monitoring Society and Social Change, GESIS Leibniz Institute for the Social
Sciences, B6 4-5., Mannheim, 68161, Germany.
Email: stephen.quinlan@gesis.org
1045730BPI0010.1177/13691481211045730The British Journal of Politics and International RelationsQuinlan
research-article2021
Original Article
Quinlan 157
That is not to say they have been ignored. Research has emerged on special elections in
Canada (Blais-Lacombe and Bodet, 2017; Loewen and Bastien, 2010), New Zealand
(Boston, 1980), the United States (Gaddie et al., 1999; Gibbs et al., 2016; Sigelman, 1981;
Smith and Brunell, 2010) and Britain (Butler, 1949; Cook and Ramsden, 1997; King, 1968;
Mughan, 1986, 1988; Norris, 1990). Comparative studies, while scarcer, also exist
(Anderson and Ward, 1996; Feigart and Norris, 1990; Norris and Feigart, 1989; Studlar and
Sigelman, 1987). Two truisms from these forays have emerged: (1) turnout in special elec-
tions is lower than in general elections, and (2) governments lose votes (and contests).
Lacunae remain, however. Most studies feature first-past-the-post contests. There has
been scant analysis of special elections under the alternative vote (AV) system, despite
Australia and Ireland using this system to fill parliamentary vacancies. In AV elections,
one seat is contested. Voters have one vote and are presented with a list of candidates,
which they rank in descending order of preference. A candidate must achieve a majority
of votes to win (quota). The counting of votes begins with the tabulation of voters’ first
choices, the first preference. A candidate who reaches or surpasses the 50% vote quota is
elected. If no candidate reaches this threshold, the candidate(s) with the fewest votes is
excluded, with the excluded candidates’ votes distributed to the voters’ next choice (lower
preferences). This process continues until a candidate secures an absolute majority. While
Australia has used AV for its special elections since 1919 and Ireland since 1922, studies
of by-elections in these two polities are minimal. And those that exist (Feigart and Norris,
1990; Norris and Feigart, 1989 – Australia; Gallagher, 1996 – Ireland), while informative,
were limited in scope and are dated. A multifaceted and contemporary analysis of AV
special elections is required.
This contribution fills this void with its focus on AV special elections in Ireland.
Ireland provides a rich testing ground as it has held 133 special elections to fill 135 vacan-
cies in the Irish parliament between 1923 and 2019. The study also allows us to tap a
unique and hereto underexplored aspect of these contests – the potential for lower prefer-
ences (transfers) to shape the outcome. Quinlan and Schwarz (2020) show these have a
decisive impact in electing about 1 in 10 candidates in single transferable vote (STV)
elections. Still, their influence in special elections has not been empirically investigated,
and outside the descriptive work of Bennett (1996) and Jesse (2000), their impact in AV
systems is under-researched. In addition, while previous studies of special elections have
been dominated by the referendum model (Tufte, 1975), which posits economic and polit-
ical conditions at by-election time are critical in accounting for government performance,
its applicability to Irish special elections is untested. While the relationship between eco-
nomic conditions and incumbent administration performance is recognised cross-nation-
ally (see, for example, Lewis-Beck, 1988; Lewis-Beck and Stegmaier, 2013), some have
suggested Ireland may be sui generis (Harrison and Marsh, 1998; although recent research
contests this – see Lewis-Beck and Quinlan, forthcoming; Leyden and Lewis-Beck,
2017). Thus, Ireland provides a compelling test case to verify this dominant model.
But this article goes beyond bringing an under-researched case to the fore. While the
referendum model has dominated, other potential mechanisms such as the surge and
decline model (Campbell, 1960), second-order motivations (Reif and Schmitt, 1980) and
the governing penalty model (Erikson, 1988) have received less attention. The impact of
candidates, prominent in the Irish voting calculus (Marsh, 2007; Marsh et al., 2008), has
been largely overlooked (an exception is the descriptive work of Norris, 1990). While
beyond two Canadian studies (Blais-Lacombe and Bodet, 2017; Loewen and Bastien,
2010), turnout has received little attention. This contribution addresses these gaps.
To continue reading
Request your trial