Government Partisanship and Electoral Accountability: The Effect of Perceived Employment Situation on Partisan Vote Switching

Published date01 September 2019
Date01 September 2019
Subject MatterArticles
Political Research Quarterly
2019, Vol. 72(3) 727 –743
© 2018 University of Utah
Article reuse guidelines:
DOI: 10.1177/1065912918804897
How do evaluations of the employment situation affect
vote choice? What are the electoral impacts of perceptions
of unemployment under different partisan persuasions of
the government? The impact of unemployment on elec-
tions is well covered in the literature. However, less is
known about differences in voter choice. The literature on
partisan voting has established that voters who are more
concerned about unemployment tend to vote for left-lean-
ing parties that promote employment (Hibbs 1979, 1982).
Individuals with greater labor market risks and vulnerabil-
ity demand for social protection and redistribution (Rehm
2016). In contrast, the literature on retrospective economic
voting suggested that voters stop supporting the govern-
ment when unemployment increases (Duch and Stevenson
2008; Erikson 1989; Warwick 1992; for a review, see
Healy and Malhorta 2013).
There have been scholarly debates about the partisan
consequences of economic downturns. Some studies sug-
gest that left parties will reap electoral benefits when
unemployment is rising, because the electorate will sup-
port them based on expectations of employment promotion
and social protection (Emmenegger, Marx, and Schraff
2015; Wright 2012). In contrast, other research notes that
the electorate will support right parties in economically
difficult times, because voters prioritize economic recov-
ery and are concerned about the increasing tax burden (Alt
1979; Stevenson 2001). Recent studies on the electoral
effects of the Great Recession in 2008 also provide con-
trasting arguments. While some studies show partisan con-
sequences for the elections held during and after the crisis
(Lindvall 2014), others indicate that primarily retrospec-
tive reward-and-punishment behaviors emerged in those
elections (Bartels 2014).
This paper goes beyond the existing literature in
important aspects. First, by analyzing panel survey data, I
explore a micro-level process of perceptions of unem-
ployment and vote switching, which most aggregate-
level studies on economic and partisan voting have
understated. I argue that the variability of the effect of
perceptions of unemployment on vote choice can be
accounted for by examining vote-switching patterns. My
804897PRQXXX10.1177/1065912918804897Political Research QuarterlyKwon
1Korea University, Seoul, South Korea
Corresponding Author:
Hyeok Yong Kwon, Department of Political Science, Korea
University, 145 Anam-ro, Seongbuk-gu, Seoul 02841, South Korea.
Government Partisanship and Electoral
Accountability: The Effect of Perceived
Employment Situation on Partisan
Vote Switching
Hyeok Yong Kwon1
What are the electoral impacts of perceptions of unemployment under different partisan persuasions of the
government? Neither the literature on retrospective economic voting nor partisan voting has provided a compelling
answer to this question. This paper addresses this puzzle by analyzing panel surveys and leveraging differences in
government partisanship in two consecutive elections. I argue that negative evaluations of the employment situation
induce voter transition to support a left-wing party under a right-wing government, but that such voter perceptions do
not affect vote choice under a left-wing government. An analysis of a voter transition, using British Election Panel Study
1992–1997 and 1997–2001, reveals findings that support my argument. My argument suggests conditional partisan
voting effects. Essentially, the effect of economic issues on vote choice is conditional on issue salience and which party
“owns” the issue, namely, the varying levels of issue salience related to government partisanship.
unemployment, political economy, economic voting, partisan voting, government partisanship
728 Political Research Quarterly 72(3)
analysis demonstrates different electoral reactions by vot-
ers based on individual evaluations of the employment
situation. Second, this study leverages differences in two
consecutive elections, where a party on the Right and one
on the Left were in power, to distinguish retrospective
economic voting and partisan issue-oriented voting.
Accordingly, this analysis examines whether partisan or
retrospective reward–punishment behavior drives the
effect of perceptions of unemployment on vote choice.
I argue that a voter’s previous vote choice and the par-
tisan persuasion of the incumbent government mediate the
effect of individual evaluations of the employment situa-
tion. Specifically, negative evaluations of the employment
situation likely induce voters to switch to a party of the
Left, with the expectation of increased promotion of
employment policies and social protection. Although this
behavioral pattern is likely more pronounced under a
right-wing government, such vote switching is less pro-
nounced under a left-wing government. I echo the idea of
Helgason and Merola (2017) that the issue of employment
insecurity entails both the government accountability
logic of economic voting and issue-based policy prefer-
ence voting. Under a right-wing government, a perception
of employment that induces leftward voting by previous
non-left voters can be conceived as both retrospective
economic and issue-based partisan voting. In contrast, a
left-wing government tends to deal with the unemploy-
ment issue more effectively and competently than a right-
wing government, decreasing the issue salience of
unemployment (Alesina and Roubini 1997). Thus, we
expect that vote switching induced by the perceptions of
unemployment is less pronounced under a left-wing
incumbent government.
Using the British Election Panel Study (BEPS), 1992–
1997 and 1997–2001, my analysis of a voter transition
model provides evidence supporting my argument. In the
1992–1997 cycle, Conservative Party supporters who
negatively evaluated the employment situation tended to
switch votes to the Labour Party, and to a lesser extent,
the Liberal Democratic Party. However, individual evalu-
ations of the employment situation did not exert any
influence on vote switching in the 2001 election, largely
because of the lower issue salience of unemployment.
These findings were consistent in the multinomial analy-
sis as well as a binary transition analysis, and the results
were robust in various model sensitivity checks. The
findings of these analyses clearly suggest that the effect
of evaluations of the employment situation is conditional
on previous vote choice and the partisan persuasion of the
The remainder of the paper is organized as follows.
The next section presents a theoretical framework of the
effect of evaluations of the employment situation on vote
switching and proposes testable hypotheses. I then
introduce the data and variables, and present the results of
the empirical analyses. The last section concludes with
several implications for the study of comparative political
behavior and democratic accountability.
Perceptions of Unemployment and
Vote Switching
This paper is concerned with whether individual evalua-
tion of the employment situation affects vote choice and its
effect on voter choice under governments with different
ideological persuasions. It is necessary to show whether
and how the perceived employment situation impacts on
voter behavior. Moreover, it is important to show that the
effect of perceptions of unemployment varies according to
the partisan persuasion of the government.
Unemployment and the Vote
How does unemployment affect electoral behavior? First,
unemployment is distinct from other economic condi-
tions as it has a more tangible influence on the lives of the
electorate. Thus, it has more economic, social, and politi-
cal consequences than other economic conditions (Blount
2002; Weatherford 1978). From the voter’s perspective,
the issue of unemployment may be distinct from an eval-
uation of the incumbent’s economic performance (cf.
Weatherford 1978, 1983). Voters consider labor market
risks such as job losses personally important, as evident
in the salient issue coverage thereof in the media during
elections (Emmenegger, Marx, and Schraff 2015). That
said, issue salience varies over time and across elections
depending on economic conditions (Singer 2013a).
Second, subjective individual evaluations of the
employment situation affect voter choice. Becher and
Donnelly (2013) have shown that economic performance
impacts vote choice and that the effect is largely driven by
voters’ subjective evaluations of the economy. That is,
individual evaluations of the economy are the causal
mechanism linking the real economy and the vote.
According to rational models, voters are more or less capa-
ble of extracting signals from macroeconomic stimuli such
as unemployment rates (Alesina and Roubini 1997). Here,
voter perceptions are a reasonably consistent approxima-
tion of an objective reality. In contrast, behavioral models
suggest that subjective evaluations can be considered a
post hoc rationalization of vote choice and partisan identi-
fication. By evoking psychological theories such as cogni-
tive dissonance theory, this view implies the issue of
endogeneity (Anderson, Mendes, and Tverdova 2004;
Evans and Andersen 2006; Hopkins 2012; Huber, Hill, and
Lenz 2012; Wlezien, Franklin, and Twiggs 1997).
According to this view, factors such as occupation, income
position, partisan identification, or ideological disposition

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT