Retrospective voting models assume that offering more information to voters about politicians performance strengthens electoral accountability. The researchers provide experimental evidence that such information decreases incumbent party support, and more findings.
What is the effect of information about an incumbent’s level of corruption on a voter’s behavior, such as support for their political party, support for the challenger, overall election turnout, and partisan attachments?
The Mexican Federal Auditor’s Office (ASF) periodically audits selected municipalities. In this study, the researchers are concerned particularly with voter reactions to audit information about spending within the Fund for Social Infrastructure (Spanish acronym: FISM) program— a large federal transfer to municipalities that is meant to fund the provision of basic services. ASF audits show how much money was spent (out of the total allocation to the municipality), whether the spending benefitted poor neighborhoods, and how much corruption occurred – measured by the number of accounting irregularities, deviations from guidelines, and misuse of resources.
This study takes advantage of this institutional arrangement and audit mechanism to test previously un-testable hypotheses about the broader effects of information on voter behavior and beliefs.
Projeto de Pesquisa:
The treatment consisted of citizen exposure to information about the use of FISM funding for public goods that was collected by the audit agency in 2007. The audits revealed that mayors did not spend all the money they received from FISM, that they did not spend all of the money in poor areas, and that in average 30% of funds were spent in a corrupt manner. The researchers show that only 10% of voters had been aware of the existence of FISM prior to their intervention.
The researchers created and distributed a leaflet that included information about FISM corruption within each municipality where the experiment was conducted. In order to separate the effects of corruption information from effects related to receiving a leaflet of any type, there was also a placebo treatment with information on the FISM that purposefully did not have information on FISM-related corruption. Units in the control group did not receive any information at all. Precincts were assigned to one of these three possible groups.
The researchers examined the following outcomes: turnout, vote share for the incumbent party and challengers, and partisan attachments.
Units and Assignment Procedure:
The researchers conducted the field experiment in twelve municipalities contained in three Mexican states —Jalisco, Morelos, and Tabasco—in 2009. The experimental units were polling precincts in audited municipalities within these states where there was an upcoming election. Polling precincts were randomly assigned to one of three groups: treatment, placebo, or control. The experimental design ensured that assignments were balanced within municipalities.
The authors find that exposing voters to corruption information results in a 1.3 percentage point decrease in turnout (or 2.5% decrease off the baseline), a .43 percentage point decrease in votes for the incumbent party, and a .86 percentage point decrease in challenger’s votes. The researchers also determine that providing information about spending and distribution (the placebo group) has no effect on vote choices. When it is revealed that corruption was high, incumbent and challenger votes decrease at higher rates.
Table 1: Effects by Level of Corruption
Incumbent Votes/Reg. voters
Challenger Votes/Reg. voters
Identification with Incumbent’s Party
Corruption (Main Effect)
The study yields similar findings on the relationship between corruption information and partisan attachments. Corruption information causes a .07 percentage point drop in identification with the incumbent party, and de-identification is stronger among voters in highly corrupt municipalities.
This study finds that corruption revelations not only affect incumbent parties, but also have an overall effect of demobilizing voters—voters who learn about corruption are less likely to vote for challengers or turn out at all. When the corruption revealed is of the highest level, voters’ already-grim view about the expected behavior of elected politicians worsens.
The study’s authors believe the effects on overall participation are evidence that voters interpret corruption information as not only pertaining to the incumbent, but as evidence of a system-wide problem. Voters do not believe challengers can avoid committing corrupt actions in an environment where corruption is found to be pervasive.