![Sample our Area Studies journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5922/TOC-Subject-Sample-2021-Area-Studies.jpg"})
ABSTRACT
Using data collected from an embedded survey administered to a nationally representative sample in Turkey, we test three common explanations for the re-election of corrupt politicians: Lack of Reliable Information, Efficiency Corruption Trade-Off, and Co-partisanship. We find that when voters receive reliable information that the incumbent mayor is corrupt, an electoral punishment (a decrease in the incumbent’s vote share in the coming elections) follows. This decrease, however, does not necessarily result in the incumbent being voted out, as we also find that many voters sufficiently value efficiency and co-partisanship in a corrupt politician to vote to re-elect.
Acknowledgments
This work was supported by the Scientific and Technological Research Council of Turkey (TÜBİTAK) under grant number 115K566. We are grateful to the editor and three anonymous reviewers for valuable suggestions. We would also like to thank Tim Lambie-Hanson for his help with the editing.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 Rundquist, Strom, and Peters, “Corrupt Politicians”; Kurer, “Why Do Voters Support”; and Anderson and Tverdova, “Corruption.” The term was coined by Kurer, “Why Do Voters Support,” 63.
2 Rundquist, Strom, and Peters, “Corrupt Politicians,” 955.
3 Throughout the paper, to prevent confusion, we refer to the voter as “she” and to the politician as “he.”
4 Using aggregated data, Brunetti and Weder, “A Free Press” find that freedom of press is negatively correlated with corruption. Turkey held a rank of 151 (out of 180 countries) on the freedom of press rankings by Reporters Without Borders in 2016, while the status of its press freedom on the Freedom House Report in the same year was Not Free. In a recent survey carried by the Social Democracy Foundation, only thirty-eight percent of the respondents expressed trust in the judiciary (the research report is available at http://sodev.org.tr/yargi-bagimsizligi.pdf).
5 Voters who accept such trade-offs are standard in political agency models, the most common theoretical approach used in rational choice models of accountability in politics, see Barro, “The Control of Politicians”; Ferejohn, “Incumbent Performance”; and Fearon, “Electoral Accountability.”
6 Winters and Weitz-Shapiro, “Lacking Information.”
7 The embedded survey method is not without its critics, though, e.g. Anduiza, Gallego, and Muñoz, “Turning a Blind Eye.”
8 The vignettes do not include information about the challengers to the incumbent in the election, as the sheer number and the variety of (the characteristics of) the challengers would require many more control groups. Furthermore, it is not clear that undertaking such costs is worth it: Muñoz, Anduiza, and Gallego, “Why Do Voters Forgive” include the quality of challengers in their vignettes, but did not find any significant effect on the voting behavior.
9 Rundquist, Strom, and Peters, “Corrupt Politicians.”
10 Anderson and Tverdova, “Corruption.”
11 Kurer, “Why Do Voters Support.”
12 Manzetti and Wilson, “Why Do Corrupt Governments.”
13 Persson, Tabellini, and Trebbi, “Electoral Rules and Corruption.”
14 For example, see Banerjee et al., “Are Poor Voters”; Winters and Weitz-Shapiro, “Lacking Information”; and Muñoz, Anduiza, and Gallego, “Why Do Voters Forgive.” Of course, other approaches, too, have been employed. For instance, to test the Lack of Reliable Information explanation, Ferraz and Finan, “Exposing Corrupt Politicians” examine whether the release of publicly available audit reports on corruption impact the outcome of mayoral elections in Brazil. The timing of the release of these reports was random, and while in some locales it was scheduled before the elections, in others the release came after. Controlling for the incumbent’s level of corruption, they find that the incumbents in the municipalities for which the report was released before the election performed worse, with the existence of local radio stations increasing the effect size.
15 Winters and Weitz-Shapiro, “Lacking Information.”
16 Muñoz, Anduiza, and Gallego, “Why Do Voters Forgive.”
17 Ibid.
18 Winters and Weitz-Shapiro, “Lacking Information.”
19 Muñoz, Anduiza, and Gallego, “Why Do Voters Forgive.”
20 Winters and Weitz-Shapiro, “Lacking Information.”
21 Çarkoğlu, “Voting Behavior in Turkey,” 160.
22 A survey by the public opinion research and consultancy company KONDA finds that in 2014, about sixty percent of the electorate in Turkey believes that (at least) most of the politicians are involved in corrupt activities, this fraction drops to forty percent among the supporters of the political party in power (the research report is available at https://konda.com.tr/tr/rapor/yolsuzluk-ve-ekonomik-durum-algisinin-secmen-tercihine-etkisi/). Combining data from several surveys, however, Çarkoğlu shows that the fraction of voters who consider corruption as the most important problem of Turkey varied significantly between 2000 and 2016, and, in general, it was not too high. See Çarkoğlu, “Ethics and Corruption,” 213.
23 Adaman and Çarkoğlu, “Social Capital” study the attitudes of businesspeople towards corruption, and Adaman, Çarkoğlu, and Şenatalar assess the households’ perception of corruption in Turkish institutions (see Adaman, Çarkoğlu, and Şenatalar, Hanehalkı gözünden). See also Kimya, “Political Economy of Corruption” for an overview of corruption in Turkey from 1950s to present.
24 Rose, “Turkish Voters and Losers’ Consent.”
25 Ibid.
26 Hazama, “Economic and Corruption Voting.”
27 Akarca and Tansel, “Voter Reaction.”
28 Çarkoğlu, “Plus ça Change.”
29 Öniş, “Turkey’s Two Elections.” Additionally, Öniş, “The Triumph of Conservative Globalism” emphasizes the role of formal and informal redistributive mechanisms, and, Esen and Gumuşcu, “Building a Competitive Authoritarian Regime” point out the role of creating a reward and punishment mechanism for the supporters and the opponents by institutional restructuring in the AKP's electoral success.
30 Yagci and Oyvat, “Partisanship, Media and The Objective Economy.”
31 Ibid., 1.
32 Çarkoğlu, “Voting Behavior in Turkey.”
33 Çarkoğlu, “Ideology or Economic Pragmatism?” Other studies, too, note the importance of the incumbent’s economic performance on voter’s decision making. For instance, using the results of an electorate tendency survey conducted recently after the economic crisis of 2001, Başlevent, Kirmanoğlu, and Şenatalar, “Empirical Investigation” show the prevalence of economic voting in the sense that unsuccessful incumbents are punished. Based on a similar survey conducted in December 2003, at a time when the economy was partially recovered, Başlevent, Kirmanoğlu, and Şenatalar, “Party Preferences” emphasize that voters take the economic performance of the government into consideration and vote not only to punish unsuccessful incumbents but also to reward successful ones. In a more recent study, using the results of electoral tendency survey conducted in February 2014, Başlevent and Kirmanoğlu, “Economic Voting” confirm the presence of economic voting in Turkey and indicate that evaluations on economic conditions are closely associated with the likelihood of voting for the incumbent party.
34 Çarkoğlu, “Economic Evaluations vs. Ideology.” Not only the salience of economic performance evaluations, but also how far back into the past the voters look while making these evaluations may be varying. Akarca and Tansel, “Economic Performance” analyze the outcomes of twenty-five local and general elections held between 1950 and 2004, and find that voters take economic performance of the government into account but not look back beyond one year. Hazama, “Economic and Corruption Voting,” on the other hand, finds that in 2014, the long-term economic success limits the relevance of the short-run economic conditions on voters’ party choice.
35 This is not an anomaly either: the support for a corrupt and inefficient incumbent is almost identical in the survey employed by Winters and Weitz-Shapiro, “Lacking Information.”
36 Winters and Weitz-Shapiro, “Lacking Information.”
37 In almost all mayoral elections in Turkey, there are more than two candidates. In the last mayoral elections held before our survey (in 2014), the average number of candidates in 81 province of Turkey was 11.2, with a standard deviation of 3.23. In the same elections, even when we only consider the candidates who received a vote share of at least ten percent, the average number of candidates was 2.5, with a standard deviation of 0.57.
38 See note 44 in Winters and Weitz-Shapiro, “Lacking Information,” 434–5.
39 For more on the possible interpretations of political representation and an extensive review of the related literature, see Stokes, Mandates and Democracy.
40 Interestingly, in Istanbul the title for the official that was responsible for several municipal services, both in the Ottoman and early Republican times, was şehremini (the trustee of the city).
41 For example, Winters and Weitz-Shapiro, “Lacking Information,” and Muñoz, Anduiza, and Gallego, “Why Do Voters Forgive.”
42 The term “perceptual screen” has been coined by Campbell et al., The American Voter. Similarly, “partisan perceptual bias” is studied by Sherrod, “Selective Perception,” Gerber and Green, “Misperceptions about Perceptual Bias,” and Zaller, The Nature and Origins of Mass Opinion, which itself may be a solution to the “cognitive dissonance” between the voter’s act (voting for a candidate) and her opinion of that candidate, see Festinger, A Theory of Cognitive Dissonance, and Mullainathan and Washington, “Sticking with Your Vote.” In the context of support for corrupt politicians, partisans’ tendency to consider corruption accusations as propaganda and “tricks” by other parties has been noted by Rundquist, Strom, and Peters, “Corrupt Politicians,” 955. Chang and Kerr, “An Insider-Outsider Theory” find that it depends on the type of the partisans: among the insiders (those either belong to the patronage networks or the same ethnic group as the corrupt politician) in Africa, the patronage insiders are knowingly tolerant of corruption due to trade-offs, while the identity insiders are less aware of it.
43 For the co-partisanship trade-off, too, there are several potential mechanisms: a voter may vote to re-elect a co-partisan corrupt politician due to his ideology or his patronage. Our survey design does not allow us to differentiate between these mechanisms. For a study exploring these differences, see Al-Ississ and Atallah, “Patronage and Ideology.”
44 Elif and Ahmet are common names in Turkey, and they do not have strong political connotations.
45 Even in the local elections, the turnout in Turkey has been high (usually above 80 percent). Therefore, even though we cannot match the characteristics of the actual voters in our sample, as we do not know who they are exactly, a sample representative of eligible voters is a good approximation to the actual voters.
46 The estimation results are available from the corresponding author upon request.
47 Winters and Weitz-Shapiro, “Lacking Information.”
48 This is available in the on-line version. Readers may also contact the corresponding author.
49 Winters and Weitz-Shapiro, “Lacking Information.”
50 Muñoz, Anduiza, and Gallego, “Why Do Voters Forgive.”
51 Winters and Weitz-Shapiro, “Lacking Information,” 422.
52 Muñoz, Anduiza, and Gallego, “Why Do Voters Forgive” also find strong support for Hypothesis 2a.
53 Or, it is possible that co-partisanship plays a role in their decision, as we discuss in Section 5.3.
54 The mechanics behind this result is as follows. In this region a corrupt and efficient politician receives a significantly higher vote share (83.8 percent) than the average vote share of an incumbent with the same description in Turkey (45.3 percent), while the vote shares of (i) an honest and efficient incumbent, (ii) an honest and inefficient incumbent, and (iii) a corrupt and inefficient incumbent in this region all are very close to their averages in Turkey.
55 Except, it seems, the higher social classes are more tolerant of corruption in Brazil, see Winters and Weitz-Shapiro, “Lacking Information.” For this study we did not define social classes, but from our survey those who are less opposed to corruption does not seem as they would fit into the higher social classes, no matter how that is defined.
56 The only exception is the voters living in Eastern and South-eastern Anatolia: here the electoral punishment for corruption is lower when the incumbent is efficient.
57 Winters and Weitz-Shapiro, “Lacking Information.”
58 Ibid.
59 Muñoz, Anduiza, and Gallego, “Why Do Voters Forgive.”
60 Winters and Weitz-Shapiro, “Lacking Information.”
61 Klasnja and Tucker, “The Economy, Corruption, and the Vote.”
62 Klasnja, Little, and Tucker, “Political Corruption Traps.”
63 Ibid.
64 But not uniformly, see Wuthrich, “An Essential Center-Periphery.” For a review, see Sayarı, “The Study of Party Politics in Turkey.”
65 Mardin, “Center-Periphery Relations.”
66 Çarkoğlu, “Voting Behavior in Turkey,” 167.
67 On the other hand, we do not have data on religiosity. The same problem exists if one wants to use the kulturkampf explanation in Kalaycıoğlu, “Kulturkampf in Turkey.”
68 As Figure 1 in Wuthrich, “An Essential Center-Periphery,” 760, illustrates, however, many cities in these three regions used to be Democrat Party strongholds in 1950, 1954 and 1957 elections, i.e. when that cleavage is assumed to be most salient.
69 Çarkoğlu, “Voting Behavior in Turkey,” 167.
Additional information
Funding
Notes on contributors
M. Kadir Doğan
M. Kadir Doğan is an Associate Professor of Economics at Faculty of Political Sciences, Ankara University. He received a B.Sc. degree in Industrial Engineering from Boğaziçi University, and both M.A. and Ph.D. in Economics from Boston University. His main research fields are political economics, game theory and information economics. His articles have appeared in leading journals in economics and political science.
Haldun Evrenk
Haldun Evrenk is Professor of Economics at Istanbul Şehir University. He earned a B.Sc. degree from Istanbul Technical University, and both M.A. and Ph.D. in Economics from (respectively) Boğaziçi University and Boston University. He has worked as a faculty member at Suffolk University in Boston and at TOBB-ETU in Ankara, and has also been a visiting researcher at Institute for Economic Development at Boston University. His research is mainly in applied game theory, focusing on problems such as politics of corruption, political competition, and voting. His work has been published in both general interest journals in economics and the leading field journals in public choice and social choice.