How Voters Respond to Presidential Assaults on Checks and Balances: Evidence from a Survey Experiment in Turkey

• Author: Aytuğ Şaşmaz,Alper H. Yagci,Daniel Ziblatt
• DOI: http://doi.org/10.1177/00104140211066216
• Published date: 01 September 2022
• Date: 01 September 2022
• Subject Matter: Articles

Article

Comparative Political Studies

2022, Vol. 55(11) 1947–1980

© The Author(s) 2022

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/00104140211066216

journals.sagepub.com/home/cps

How Voters Respond to

Presidential Assaults on

Checks and Balances:

Evidence from a Survey

Experiment in Turkey

Aytu˘

gS¸as¸maz

1

, Alper H. Yagci

2

, and Daniel Ziblatt

3

Abstract

Why do voters support executive aggrandizement? One possible answer is

that they do so because they think this will ease their preferred leader’s hand

in putting their partisan vision into action, provided that the leader will

continue winning elections. We study this phenomenon through a survey

experiment in Turkey, by manipulating voters’perceptions about the po-

tential results of the f‌irst presidential election after a constitutional refer-

endum of executive aggrandizement. We f‌ind that voters from both sides

display what we call “elastic support”for executive aggrandizement; that is,

they change previously revealed constitutional preferences in response to

varying winning chances. This elasticity increases not only when citizens feel

greater social distance to perceived political “others”(i.e., affective polari-

zation) but also when voters are concerned about economic management in a

potential post-incumbent era. Our f‌indings contribute to the literature on

how polarization and economic anxiety contribute to executive aggran-

dizement and democratic backsliding.

1

1

Stanford University, Stanford, CA, USA

2

Bogazici University, Istanbul, Turkey

3

Harvard University, Cambridge, MA, USA

Corresponding Author:

Alper H. Yagci, Department of Political Science and International Relations, Bogazici University,

Bebek, Istanbul 34342, Turkey.

Email: alper.yagci@boun.edu.tr

Keywords

democratic backsliding, executive aggrandizement, referenda, voter behavior,

polarization

Democratic backsliding frequently occurs today through executive

aggrandizement—elected political executives dismantling the constraints of

constitutional systems of checks and balances (Bermeo, 2016). The path to

executive aggrandizement is often paved as incumbents attempt to garner

voters’approval in referenda for such initiatives (Graham & Svolik, 2020;

Svolik 2019;Levitsky & Ziblatt, 2018). But why would voters support in-

stitutional changes that make the incumbent less accountable and constrained?

One frequently hypothesized explanation is that voters do so when they

think that it could be good for their side. Aspiring autocrats rely on their

partisan supporters’belief that any move away from democratic constraints

will ease his hand in putting their partisan policy vision into action and

entrench their dominance in society. But major constitutional institutions are

sticky and have effects that are lasting beyond a single electoral cycle. In

settings where there is some electoral uncertainty, buying into the autocrat’s

vision entails a risk for the voters because once put in place, authoritarian

institutions can be used against them if and when an opponent wins elections

and comes to power. Generating an aura of invincibility and reducing electoral

uncertainty is thus useful for would-be autocrats because this assures their

followers that lending support to executive aggrandizement can only serve to

strengthen the place of their preferred policies in the public policy mix and/or

buttress their political dominance.

This idea is intuitive, but has not been subject to much empirical testing. In

this study, we present a test relying on a population-based experiment where

we artif‌icially manipulate electoral uncertainty and examine how voters’

support for executive aggrandizement changes in response. We also explore

the attitudes that induce people to change their constitutional choices op-

portunistically in response to varying electoral uncertainty. When there is a

high level of partisan polarization, the stakes are higher for the decision of

endorsing authoritarianism—because a “wrong”electoral outcome can lead to

the strengthening of a highly undesired rival vision. Similarly, when voters

regard the incumbent as uniquely capable in policy management, their choices

on how to allocate political power would respond to variations in electoral

certainty. For people whose opinions on positional or valence considerations

are more distant from the other side, then, the variation in expected election

outcomes should be a more important determinant of constitutional choice.

In this study, we examine the phenomenon of popular support for executive

aggrandizement in Turkey, a quintessential case of democratic backsliding in

1948 Comparative Political Studies 55(11)

the contemporary era (Bermeo, 2016). In 2017, Turkey held a national citizen

referendum on whether to shift from a semi-presidential to an autocratic super-

presidential system by removing the prime minister’soff‌ice and signif‌icantly

weakening checks and balances on presidential executive power. The in-

cumbent president Tayyip Erdo˘

gan and his party AKP campaigned for a “yes”

vote and the main opposition parties did the opposite. Existing research shows

that the voters’behavior in the referendum closely followed their partisan

identities (Aytaç et al., 2017). Ayear late r in 2018, however, in the run-up to

the f‌irst presidential election of the new regime, opposition candidates seemed

to put on a surprisingly powerful campaign for the presidency. Weutilized this

setting to administer a survey experiment a week before the presidential

election, where respondents were given varying information about the re-

election chances of the incumbent candidate. We explored whether voters

changed their previously revealed views in support of, or against, executive

aggrandizement upon receiving new information about their preferred can-

didate’s chances of winning. To preview our f‌indings, despite our weak

treatment, at least 10% of voters have elastic support for executive aggran-

dizement, that is, they change their institutional preferences opportunistically

upon new information. This attitude is existent in both pro-government and

anti-government camps (some of whom start supporting a previously rejected

authoritarian setup upon updated winning chances), though in different

degrees.

In addition, we explored competing logics that could push voters to shift

their institutional preferences opportunistically on the basis of their preferred

candidate’s winning prospects: The tendency to change views may hinge on

levels of affective positional polarization; or, it may depend on a valence

judgment about how the important of their candidate for economic man-

agement. While these are not mutually exclusive hypotheses, they correspond

to two rival models of electoral authoritarian legitimacy—those that em-

phasize the degree to which political leaders exploit affective partisan po-

larization to bolster their power (Svolik 2019,2020;Nalepa et al., 2018) and

those that emphasize the degree to which “informational autocrats”manage to

convince the public of their economic management competence (Guriev &

Treisman, 2019). If the f‌irst consideration is strong, the power-seeking in-

cumbent could f‌ind it useful to increase polarization by appealing to the

desires and fears of one side only. If, on the other hand, voters are most swayed

to support an incumbent by arguments about economic competence of the

executive, the incumbent would have to form a broader appeal and deliver on

economic performance. We f‌ind evidence of both motivations, though the

affective polarization dynamic is more important.

There is a growing literature on the effects of heightened partisan po-

larization. Our contribution to this debate is an analysis of its implications for

regime change: if partisan reasoning extends to voters’preferences on the

S¸as

¸maz et al. 1949