close× Call Us
close×
Title Does Corruption Signal Competence in Politics? Experimental Evidence from India
Post date 04/06/2019
C1 Background and Explanation of Rationale

Social scientists have found that corruption impedes economic growth and democratization. However, there is abundant evidence that voters across a wide range of contexts support politicians who have been accused of corruption. Why? In this paper, we propose and test the hypothesis that corruption can serve as a positive signal of a politician's perceived competence, or ability to ``get things done.'' Our hypothesis is thus distinct from existing explanations centering on voter ignorance or the willingness of voters to ``forgive'' politicians' corruption should they perform well. We conduct a survey experiment with over 1,000 respondents in Maharashtra state, India, wherein subjects are randomly assigned a vignette describing two candidates running for statewide office and asked how they would vote. Our design allows us to compare voter support, relative to a honest politician, for three types of candidates for office: a corrupt candidate, and corrupt and honest candidates whose past performance indicates average competence. We theorize that if corrupt politicians are viewed by voters as having above-average competence, then information indicating average competence should lead voters to update downward and to decrease their support. In contrast, the same information about a honest politician should lead voters to update upward and increase their support. To evaluate our hypotheses, we will supplement vignette voting outcome data with data on voters' perceptions of politician competence. Finally, we will consider under what conditions we should expect effects to be greatest: for instance, for those voters who believe that the country faces intractable problems or who believe that corruption is widespread in Indian politics.

C2 What are the hypotheses to be tested?

H1: Respondents will prefer honest politicians to corrupt politicians.
H2: Respondents will prefer corrupt candidates about whom no information on competence is provided to corrupt candidates whose past performance indicates average competence.
H3: Respondents will prefer honest candidates whose past performance indicates average competence to honest candidates about whom no information on competence is provided.
H4: Perceptions of politicians' competence will fall in the following order (from high to low): HI>CNI>HNI>CI, where C refers to a corrupt politician, H a honest politician, I means information indicating past average competence is provided, and NI indicating no information about past average competence is provided.
H5: Effects predicted by H2-H4 should be increasing in an individual's stated belief that corruption is widespread and necessary to succeed in politics, the country faces intractable problems, and corrupt politicians are more competent than honest politicians; valuation of competence over ideology, delivery of benefits over shared ethnic identity or honesty, and shared ethnic identity over honesty; and their prior experience voting.

C3 How will these hypotheses be tested? *

See Table 2 in the PAP for the hypotheses and how they are to be tested. Respondents will be assigned to three different experimental treatments to evaluate our hypotheses. We will conduct difference in means and proportions tests, will conduct subset analyses, and will also run some probit and ordered probit regression analyses.

C4 Country India
C5 Scale (# of Units) Over 1,000
C6 Was a power analysis conducted prior to data collection? Yes
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number Approved by New York University's Institutional Review Board (IRB) under protocol #IRB-FY2019-2859, by Centre for Experimental Social Sciences (CESS) Nuffield's IRB under protocol #FE_0015, and by FLAME University's IRB under protocol #2019-02-001.
C9 Date of IRB Approval NYU (February 19, 2019), Flame University (February 18, 2019), CESS Nuffield (February 27, 2019)
C10 Will the intervention be implemented by the researcher or a third party? Morsel Research & Development Pvt Ltd
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research? No
C12 If relevant, is there an advance agreement with the implementation group that all results can be published? Yes
C13 JEL Classification(s) D73, D72