Group-Specific Responses to Retrospective Economic Performance: A Multilevel Analysis of Parliamentary Elections

• Author: Abel Bojar,Tim Vlandas
• Published date: 01 December 2021
• Date: 01 December 2021
• DOI: http://doi.org/10.1177/0032329221989150
• Subject Matter: Articles

https://doi.org/10.1177/0032329221989150

Politics & Society

2021, Vol. 49(4) 517 –548

© The Author(s) 2021

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0032329221989150

journals.sagepub.com/home/pas

Article

Group-Specific Responses

to Retrospective Economic

Performance: A Multilevel

Analysis of Parliamentary

Elections

Abel Bojar

European University Institute

Tim Vlandas

University of Oxford

Abstract

What is the relationship between electoral and economic performance? Previous

literature posits that poor economic performance hurts the incumbent at the ballot

box because overall economic performance serves as a competence signal, which

voters can readily access at low costs. Building on an emerging economic voting

literature exploring heterogeneity in the electorate, this article argues that social

groups are affected differently by various dimensions of economic performance and

that their sociotropic sanctioning of incumbents is contingent on the retrospective

performance of these dimensions. It theorizes how four social groups—low-skilled

workers, pensioners, public sector employees, and high-income individuals—are

differently affected by each of four economic dimensions: unemployment, inflation,

stock market performance, and public spending; as a result, they penalize the

incumbent to varying extents. Results from a multilevel logistic regression analysis

from four modules of the Comparative Study of Electoral Systems containing around

seventy electoral contexts are consistent with the argument.

Keywords

economic voting, unemployment, inflation, public spending, stock market, political

economy

Corresponding Author:

Abel Bojar, European University Institute, Badia Fiesolana via dei Roccettini, 9, 50014 San Domenico di

Fiesole, Florence, Italy.

Email: Abel.Bojar@eui.eu

989150PASXXX10.1177/0032329221989150Politics & SocietyBojar and Vlandas

research-article2021

518 Politics & Society 49(4)

Economic voting postulates a simple formula of electoral accountability: deteriorat-

ing economic performance has a negative impact on the reelection chances of

incumbents because the poor competence of the government that it signals is penal-

ized by the electorate (sociotropic channel) or because it affects voters negatively

(egotropic channel).1 However, this expectation is subject to conflicting results in

the literature: the impact of the economy on electoral support seems to vary across

time, countries, and contexts. Some authors address the puzzle by pointing to the

mediating role of institutions via clarity of responsibility and the welfare state, while

others instead argue that the economic vote is contingent on international economic

dynamics.2

Much of the literature conceptualizes retrospective economic performance only as

a competence signal regardless of group-specific preferences and risk exposures:3 if

the economic vote is used as a “shortcut” to evaluate government competence without

paying the costs of being informed about all government policies in different domains,

the economy can be considered a valence issue that roughly captures incumbents’

competence, and the effect of unemployment, inflation, or economic growth on elec-

toral support is therefore largely independent of the socioeconomic status of the

respondent.

An emerging economic voting literature instead explores heterogeneity within the

electorate and challenges the assumption that all voters can be thought of as penalizing

incompetence to a similar extent and that they constitute a more or less homogenous

group deriving similar utilities from variations in economic performance. Some stud-

ies have empirically tested for the differential impact of the economy along income,

class, and asset ownership;4 others have focused on heterogeneous accountability due

to information, self-interest and corruption scandals, or senior political activism and

the welfare state.5

In this article, we build on this emerging literature to further explore heterogeneity

among the electorate in the way voters are affected by—and react to—changes in dif-

ferent economic dimensions. We propose an overarching theoretical framework based

on the idea of group-specific economic voting that allows us to model how distinct

electoral groups may be affected differently by various aspects of the economy. In

doing so, we broaden the scope of economic voting research by investigating how the

electoral impact of the economy varies by social groups not just quantitatively (i.e.,

how much does the economy matter) but also qualitatively (i.e., what aspect of the

economy matters most for whom).

Moreover, we wish to strengthen the microfoundations for key insights in the com-

parative political economy (CPE) literature by zooming in on the electoral mechanism

underpinning partisan politics. Previous studies have shown that different social

groups have different risk profiles and distinct economic and welfare state policy pref-

erences, which have important implications for electoral behavior.6 For instance, in the

domain of labor market policy, social democrats rely increasingly on the electoral sup-

port of labor market insiders in permanent, well-protected contracts and as a result

may not pursue policies that benefit outsiders.7 Similarly, the expected electoral

response to austerity policies is heavily contingent on individuals’ socioeconomic

Bojar and Vlandas 519

resources to withstand cutbacks in government programs.8 In this article, we directly

investigate the electoral mechanism that is typically assumed, but not tested, by most

existing CPE studies.

We develop a systematic account of how different aspects of the economy affect

the electoral behavior of different social groups. In addition to the usual suspects of

unemployment and inflation as the main objects of responsibility attribution, we

also integrate novel aspects of retrospective economic performance. First, with

increasing financialization and inequality, asset market performance marks an

important demarcation line between well-off savers with financial (and nonfinan-

cial) assets and lower-income wage earners who stand to gain little from financial

booms.9 Second, over the last few decades, budgetary policy choices with highly

politicized austerity packages have brought back economic policy to the forefront of

political debates, leading us to include fiscal policy measures as a fourth dimension

in our analysis.10

We derive several observable implications of our argument with respect to the vot-

ing behavior of several social groups along their different skill profiles, labor market

status, income, and sector of employment, which all entail different exposures to

unemployment, inflation, public spending, and stock market returns. We then test

these observable implications using multilevel mixed-effects logistic regression analy-

sis of three modules of the Comparative Study of Electoral Systems (Modules 2–4).

Consistent with our expectations, we find that the electoral impact of deteriorating

economic performance is largely contingent on individuals’ risk profiles, with pen-

sioners appearing particularly responsive to retrospective inflation performance, low-

skilled workers to unemployment levels, public sector workers and low-skilled

workers to public sector cuts, and high-income individuals to developments in the

stock market.

The central concept of our contribution—individual-level heterogeneity behind

the electoral response to the economy—builds on well-established findings in the

political behavior literature. In addition to the material self-interest logic underpin-

ning our approach, other studies have documented further sources of differential

reactions to the economy, such as political sophistication and partisanship and insti-

tutional trust.11 It is possible that electoral contexts greatly differ in the extent to

which other macro-level considerations, such as important events, dominate the

political agenda, thereby rendering the salience of the economy secondary.12

Although a full account of all potential confounders lies beyond the scope of our

contribution, we control for a number of individual-level variables besides the socio-

economic groups that constitute the central focus of our study to address this limita-

tion. Moreover, we are relatively sanguine about the noneconomic sources of

political support at the macro level because we have no reason to presuppose that

they systematically covary with prior economic conditions in a way that would bias

our coefficient estimates.

The remainder of this article proceeds as follows. In the next section, we review the

existing literature and further substantiate its shortcomings. We then develop a model

of economic voting that takes into account the different risk profiles of voters. The