Economic Perception to Political Performance Evaluation: Establishing Precursors to Economic Voting in Africa

• Author: Inbok Rhee
• DOI: 10.1177/1065912919888017
• Published date: 01 March 2021
• Date: 01 March 2021
• Subject Matter: Articles

https://doi.org/10.1177/1065912919888017

Political Research Quarterly

2021, Vol. 74(1) 131 –147

© 2019 University of Utah

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/1065912919888017

journals.sagepub.com/home/prq

Article

Introduction

Established theories of economic voting predict that vot-

ers can use elections to either reward or punish incumbent

politicians by engaging in a retrospective evaluation of

economic performance. Existing studies on the economic

vote find consistent and robust support in Western

Europe, the United States, Latin America, and Eastern

Europe (Lewis-Beck and Stegmaier 2013). Yet, observers

have often been skeptical about the role of elections in

rewarding or punishing politicians in African democra-

cies, due to factors such as the relatively short history of

African democracies, prevalence of dominant parties,

and clientelistic and ethnic appeals. Empirical evidence

has also been weak: the effects of economic factors on

voter behavior—when found—are weaker than those of

noneconomic factors, such as ethnicity or geography

(Bratton, Bhavnani, and Chen 2012; Posner and Simon

2002; Youde 2005).

This paper investigates economic voting in sub-Saha-

ran Africa by exploiting new theoretical indicators and

using both subjective economic evaluation and objective

economic indicators to capture economic changes.

Economic voting involves a long causal chain consisting

of at least four components: changes in objective

economy, subjective economic perception, subjective

politician performance evaluation, and the eventual polit-

ical action at the polling booth (Anderson 2007; Becher

and Donnelly 2013; Healy and Malhotra 2013). Relying

only on subjective economic perception as the key inde-

pendent variable can make inference difficult, as eco-

nomic perceptions can be a consequence of incumbent

support rather than an influence on it (Bartels 2002;

Evans and Andersen 2006; Evans and Pickup 2010).

Employing only objective economic indicators is also

problematic as changes in the local economy can condi-

tion the influence of any changes at the national-level

macroeconomic indicators on political perception or

behavior (Ansolabehere, Meredith, and Snowberg 2011;

Reeves and Gimpel 2012).

Using more than 55,000 individual-level geo-located

observations across sixteen countries in sub-Saharan

Africa, I test the association between subjective economic

888017PRQXXX10.1177/1065912919888017Political Research QuarterlyRhee

research-article2019

1KDI School of Public Policy and Management, Sejong, Republic of

Korea

Corresponding Author:

Inbok Rhee, KDI School of Public Policy and Management, 263

Namsejong-ro, Sejong 30149, Republic of Korea.

Email: inbok.rhee@kdis.ac.kr

Economic Perception to

Political Performance Evaluation:

Establishing Precursors to

Economic Voting in Africa

Inbok Rhee1

Abstract

Empirical support for economic voting is well documented in advanced democracies. We know less, however, about

the extent and dynamics of economic voting in the developing democracies of sub-Saharan Africa. The relationship

between economic perceptions and incumbent performance evaluations is a critical precursor to vote choice. I

evaluate this link using more than fifty-five thousand individual-level observations across sixteen sub-Saharan African

countries. I find that there exists a strong association between economic perception and performance evaluation

while controlling for a host of covariates, including ethnicity, partisanship, information, and public goods provision.

Contrary to previous findings, however, I show that the influence of economic perception is stronger than many

other factors considered in the models such as coethnicity with the incumbent. Moreover, my findings indicate that

coethnicity—but not copartisanship—conditions the influence of economic perception on performance evaluation. I

use an instrumental variables approach to further validate the findings.

Keywords

economic voting, Africa, accountability, public opinion

132 Political Research Quarterly 74(1)

perception and performance evaluation. To preview the

results, I find a strong and consistent association between

the two. Such association is robust to the inclusion of a

battery of covariates, including socioeconomic status

(SES), electoral proximity, levels of information, public

goods provision, ethnicity, and partisanship, to different

model specification, and to additional sensitivity analysis.

The results also show that subjective economic percep-

tions have the strongest association on incumbent perfor-

mance evaluations. This effect, however, is conditioned

by ethnicity—but not by partisanship. While both coeth-

nics and non-coethnics of the executive reward good eco-

nomic performance, coethnics are more forgiving of poor

economic performance. Finally, I confirm the main results

using additional robustness checks including an instru-

mental variables (IV) approach exploiting the objective

changes in the economy to instrument for the subjective

economic perception.

The study makes a number of contributions. First, there

is limited understanding about the prevalence of economic

voting in sub-Saharan Africa. The complexities of African

societies, especially ethnic cleavages, economic inequality,

and different levels of economic development provide a

real challenge to economic voting research. This paper con-

siders these societal and economic characteristics, and tests

for both the independent and conditional associations of

partisanship and ethnicity in the context of economic vot-

ing. It also advances the literature that examines factors

affecting the magnitude of economic voting, such as gov-

ernment structures and economic trade (Hellwig and

Samuels 2007; Nadeau, Niemi, and Yoshinaka 2002).

Second, this paper parses out the different steps required for

economic voting and focuses on one segment of the eco-

nomic voting process, namely, the connection between sub-

jective economic evaluation and subjective political

performance evaluation. This helps us better understand an

important step before the voting act—namely, political per-

formance evaluation—which is also crucial for govern-

ment’s legitimacy, ability to enact policies, and ability to

engage in the international community. This also allows me

to bring in the changes in the objective economic conditions

to instrument for the subjective economic perception.

Finally, many existing studies of the topic focus on single-

country case (Arriola 2007; Bratton and Kimenyi 2008;

Ferree 2006; Gibson and Long 2009; Hoffman and Long

2013; Lindberg and Morrison 2008; Posner and Simon

2002; Youde 2005), whereas this paper makes a systematic

cross-national model of economic voting (Bratton,

Bhavnani, and Chen 2012; Norris and Mattes 2003).

Theoretical Foundation

Consistent with the sanctioning model of electoral com-

petition (Barro 1973; Ferejohn 1986), scholars found a

strong correlation between changes in economy and vot-

ing. Prior works have shown that voters are more influ-

enced sociotropically than pocketbook calculations

(Kiewiet 1983; Kinder and Kiewiet 1979), weigh the ret-

rospective evaluation of incumbent performance more

than prospective expectations (Fiorina 1981; Key 1966),

and are possibly harsher in terms of punishments than

rewards (Bloom and Price 1975). These findings are not

exclusive—for instance, voters are both retrospective and

prospective. But the idea of retrospective and sociotropic

voting has been the dominant model.

While much of the early empirics on economic voting

focused on elections in Western Europe and the United

States (Lewis-Beck 1988; Lewis-Beck and Paldam 2000;

Markus 1988), there have also been efforts to examine

similar relationships in developing democracies, espe-

cially in Latin America and Eastern Europe (Lewis-Beck

and Ratto 2013; Lewis-Beck and Stegmaier 2008; Tucker

2002). Although there are some variations in terms of sta-

tistically and substantively significant effects across stud-

ies conducted in different contexts, the notion that

governments are held accountable for bad economic

management seems to hold with few exceptions (Lewis-

Beck and Stegmaier 2007).

Economic Voting in Sub-Saharan Africa

In the context of African politics, observers have doubted

the link between retrospective economic evaluation and

voting for a number of reasons (Hoffman and Long 2013).

First, the history of electoral democracy is relatively

young, and attribution of responsibility may not function

as well as in advanced democracies (Van de Walle 2003).

Second, voters may have difficulties in assessing politi-

cian performance given the low information environment

(Posner 2005). Third, the typical party system consists of

a dominant party surrounded by a large number of small

and unstable parties, which discourages opposition votes,

and gives incumbent parties dominant positions (Van de

Walle 2003). Fourth, electoral contests are often easily

dominated by valence appeals instead of substantive pol-

icy positions (Bleck and Van de Walle 2011, 2013).

Finally, strong executives have arguably enjoyed signifi-

cant electoral advantages due to factors ranging from

name recognition, access to state resources, and popular

aversion to voting against incumbents (Bratton and Logan

2015; Cho and Logan 2014).

Some recent scholarship challenges these views and

shows that performance also matters to citizens in

Africa. These studies can be classified into four catego-

ries depending on their selection of variables. First,

studies relying on survey data have often tested the

effects of subjective evaluation of the incumbent perfor-

mance on self-reported vote choice or intent to vote.