1. Introduction
Candidates and political parties across the ideological spectrum adopt anti-corruption rhetoric into their electoral campaigns. While some candidates genuiniely commit to substantial reform, others merely aim to capitalize on public frustration with the political establishment. In fact, the anti-corruption agenda is often captured by populist leaders and politicians who themselves face allegations or investigations of criminal offenses. By promising to clean government and change the way politics is conducted, they deflect criticism for their own criminal offenses and portrait their opponents as disconnected elites who embody the problems of systemic corruption. This opportunistic adoption of anti-corruption rhetoric raises a paradox: while advocating for swift action against corruption can pave the way for meaningful reform, anti-corruption platforms can also allow corrupt politicians to escape significant electoral consequences. If anti-corruption rhetoric can protect corrupt politicians from scrutiny or help populist candidates win elections without leading to meaningful reforms, it is crucial that we deepen our understanding of how the public responds to these appeals during campaigns.
This paper addresses two critical questions regarding the dynamics of anti-corruption campaigns. First, is campaigning on anti-corruption platforms effective in attracting voter support? Second, under what conditions is anti-corruption rhetoric more persuasive? Addressing these questions is important because the appeal of anti-corruption messages might clarify why many corrupt politicians often escape accountability or why insincere reformists frequently win elections only to obstruct reform later. In order to elucidate the appeal of anti-corruption rhetoric, this study will systematically examine various scenarios in which politicians adopt anti-corruption platforms. Specifically, the analysis focuses on how the content of the message—whether it presents a policy-specific proposal—and the profile of the messenger—particularly their history of corruption and their gender—influence the effectiveness of anti-corruption rhetoric.
To the best of my knowledge, this is the first political science study to specifically explore the impact of anti-corruption appeals. Previous research has thoroughly investigated how being implicated in corruption investigations and prosecutions undermines candidates’ electoral prospects (Anduiza et al. Reference Anduiza, Gallego and Muñoz2013; Botero et al. Reference Botero, Cornejo, Gamboa, Pavao and Nickerson2015; De Vries and Solaz Reference De Vries and Solaz2017; Eggers Reference Eggers2014; Klašnja et al. Reference Klašnja, Lupu and Tucker2021). However, the strategic use of anti-corruption rhetoric—a tactic employed across the ideological spectrum by both populist and non-populist candidates in varying corruption environments, including by candidates accused of corruption—has been largely neglected. Given the prominence of anti-corruption appeals, understanding the consequences of politicizing corruption requires examining not only the electoral fate of those accused of wrongdoings but also the effectiveness of anti-corruption promises. This study provides a systematic assessment of how anti-corruption messages influence electoral outcomes, aiming to bridge this gap and offer new insights into the dynamics of corruption politicization and electoral accountability.
To study anti-corruption platforms, I designed an unforced conjoint experiment embedded in a survey in Paraguay, a Latin American country with high levels of corruption (Albisu Ardigó Reference Albisu Ardigó2016). While the majority of corruption experiments have been conducted in relatively prosperous Latin American countries (Batista Pereira Reference Batista Pereira2021; Botero et al. Reference Botero, Cornejo, Gamboa, Pavao and Nickerson2015, Reference Botero, Castro Cornejo, Gamboa, Pavao and Nickerson2019; Chong et al. Reference Chong, De La, Karlan and Wantchekon2015; Weitz-Shapiro and Winters Reference Weitz-Shapiro and Winters2017; Winters and Weitz-Shapiro Reference Winters and Weitz-Shapiro2018), the less affluent country of Paraguay has been overlooked despite possessing a deep-rooted history of machine politics. The enduring nature of corruption and the resilience of its dominant party (the Colorado Party) makes Paraguay a case where garnering public support for anti-corruption measures is expected to be challenging. While citizens might generally support efforts against corruption, Paraguayans are likely to be highly skeptical and cynical about the actual potential for meaningful change. Paraguay has had an active public agenda focused on anticorruption reforms, yet often these initiatives, like increasing government transparency, have suffered from weak enforcement and limited impact. Failed anticorruption reforms coupled with the normalization of corruption may foster public distrust in anticorruption pledges. The experimental design varies several features of the hypothetical candidate running for mayor, including anti-corruption platform, gender, and disciplinary record. After each pair of profiles, respondents answer two questions: which candidate they would vote for, and which candidate they believe will help reducing corruption the most. Having these two outcomes allows me to assess the effectiveness of anti-corruption appeals in winning votes and their ability to persuade citizens of a candidate’s commitment and effectiveness in combating corruption.
The results show that, on average, anti-corruption appeals have a significant effect on electoral support. Candidates who advance an anti-corruption platform are more likely to win electoral support than candidates who do not pledge to fight corruption. I also find that concrete anti-corruption platforms with specific policy proposals are more persuasive of a candidate’s commitment and effectiveness in fighting corruption than vague anti-corruption platforms. Finally, the results show that the credibility of the candidate influences how compelling anti-corruption pledges are, although it does in unexpected ways. Having a clean corruption record does not make anti-corruption pledges more persuasive, suggesting that citizens do not use misconduct to assess the trustworthiness of politicians as anti-corruption crusaders. Additionally, being a woman does not make anti-corruption candidates more persuasive than their male counterparts. On the contrary, women’s credibility advantage disappears when candidates adopt anti-corruption appels, suggesting that male candidates gain an electoral advantage from these platforms that women do not.
These findings have important implications for our understanding of the appeal of anti-corruption pledges and electoral accountability. The present study shows that isolated anti-corruption promises, even when detached from the populist framework, can independently increase electoral support. Promises to clean government can resonate with voters whether they are vague or concrete, though specific appeals accompanied by substantive plans are more persuasive of a candidate’s effectiveness in fighting corruption. This suggests that candidates are incentivized to adopt electoral appeals with policy content, potentially enhancing accountability by providing voters with tools to assess the future performance of politicians. These findings not only hold potential normative implications for theories of representative democracy, but they also shed light on an overlooked aspect of electoral accountability for corruption. Specifically, the results show that anti-corruption rhetoric provides a similar electoral reward for corrupt and clean candidates, which can help explain why corrupt politicians often escape accountability. When corruption is seen as a normalized practice, citizens may be inclined to grant corrupt politicians the same credibility as clean ones. Finally, this study also contributes to the literature on gender and corruption. The finding that vague anti-corruption platforms can harm female candidates’ electoral support adds nuance to the notion that women in politics are held to higher standards.
2. Corruption, Anticorruption, and Voting Behavior
The conventional wisdom about corruption and voting behavior is that citizens punish incumbents for corruption at the ballot box only under certain conditions. A growing literature has explored the factors under which candidates accused of corruption are penalized by voters. Several hurdles to electoral accountability for corruption have been identified, such as: partisan loyalties (Anduiza et al. Reference Anduiza, Gallego and Muñoz2013; Eggers Reference Eggers2014), the economy and side-benefits (Klašnja and Tucker Reference Klašnja and Tucker2013; Klašnja et al. Reference Klašnja, Lupu and Tucker2021), unreliable information (Weitz-Shapiro and Winters Reference Weitz-Shapiro and Winters2017; Winters and Weitz-Shapiro Reference Winters and Weitz-Shapiro2018), widespread corruption and the lack of clean or viable alternatives (Agerberg Reference Agerberg2020; Bauhr and Charron Reference Bauhr and Charron2018; Elia and Schwindt-Bayer Reference Elia and Schwindt-Bayer2021; Pavão 2018; Vera Reference Vera2022), among others.
However, the existing literature has predominantly focused on studying the incumbents affected by the politicization of corruption —those who risk losing their post due to rising corruption under their watch or corruption scandals directly involving their administration. This focus has resulted in a neglect of examining the electoral appeal of the beneficiaries of the anti-corruption rhetoric —the candidates who capitalize on citizens’ concerns regarding rising corruption. These anti-corruption candidates often accuse their opponents of corruption and propose reforms to combat it (Bågenholm Reference Bågenholm2013), reglardless of their own questionable records. They adeptly leverage the context of corruption scandals surrounding the incumbent government, emphasizing the need to fight corruption in their campaigns. Nevertheless, our knowledge is primarily confined to the electoral impact of corruption allegations, rather than the influence and persuasiveness of the anti-corruption rhetoric.
This research gap prompts important questions: Are anti-corruption promises always effective? Can citizens discern between mere rhetoric and credible commitments? Who is considered a more credible anti-corruption candidate? While existing literature has focused on the factors that mitigate electoral sanctions for corrupt politicians, it has overlooked the relevance of anti-corruption pledges. To gain a deeper understanding of the conditions under which anti-corruption promises are effective, we need a theory that explains which aspects of the anti-corruption message and qualities of the messenger resonate with voters and compel their support. The overall aim of this paper, therefore, is to uncover the underlying factors that contribute to the effectiveness and persuasiveness of the anti-corruption platform. This would provide valuable insights into the factors that reduce electoral penalties for corrupt politicians and deepen our understanding of the broader phenomenon of leveraging anti-corruption appeals in political campaigns.
Prior anti-corruption research primarily focused on strategies and policies for controlling corruption (Gans-Morse et al. Reference Gans-Morse, Borges, Makarin, Mannah-Blankson, Nickow and Zhang2018; Rose-Ackerman Reference Rose-Ackerman1999). These studies have explored the effectiveness of various reforms and approaches, including monitoring and oversight institutions, transparency, and campaign finance reform (Dávid-Barrett and Fazekas Reference Dávid-Barrett and Fazekas2020; Hummel et al. Reference Hummel, Gerring and Burt2021). While this research has advanced our understanding of implementation and success factors (Mungiu-Pippidi Reference Mungiu-Pippidi2015; Persson et al. Reference Persson, Rothstein and Teorell2013), it has often overlooked the electoral dimension of anti-corruption proposals. Consequently, what renders anti-corruption candidates’ messages persuasive remains underexplored. The present study aims to illuminate the appeal of anti-corruption promises in electoral campaigns.
This article advances the literature by developing a theory of the conditions under which anti-corruption appeals become effective. There is considerable theoretical ambiguity regarding the effectiveness of anti-corruption promises. While some voters may find anti-corruption promises appealing and view the candidates as genuine reformers, others may approach them with skepticism, considering them mere political rhetoric aimed at securing votes. Nevertheless, corruption is widely disliked by voters, as it undermines trust in government institutions. By tapping into this sentiment, anti-corruption candidates can connect with voters who are eager for change. The next section investigates the appeal of anti-corruption rhetoric.
3. Are Anticorruption Promises Effective?
There are theoretical reasons to suspect that anti-corruption platforms can help candidates win votes, especially in contexts where corruption is a public concern. First, electoral promises to fight corruption can be effective in persuading citizens and winning elections because they resonate strongly with voters’ dislike of corruption, and evidence suggests citizens in the Global South hold strong anti-corruption norms (Boas et al. Reference Boas, Hidalgo and Melo2019). Voters strongly oppose corruption as it represents a betrayal of the trust that citizens place on public officials to act in the best interest of the people (Rose-Ackerman Reference Rose-Ackerman2001). Public opinion surveys reveal strong aversion to public official’s misconduct and a demand for clean government in Europe, Latin America and the United States. A majority of respondents worldwide report that most or almost all public officials in their country are involved in corruption (Inglehart et al. Reference Inglehart, Haerpfer, Moreno, Welzel, Kizilova, Diez-Medrano, Lagos and Norris2022). Fifty seven percent of people globally think the government is not doing enough to fight corruption (Global Corruption Barometer 2017). Citizens concerned about clean government often express a desire for transparent and accountable government where public officials act in the best interest of the public and maintain the highest standards of integrity. As rational parties and candidates aim to associate themselves with winning issues (Hobolt and de Vries Reference Hobolt and de Vries2015), embracing anti-corruption measures becomes an effective electoral strategy for maximizing votes, especially where corruption is perceived to be on the rise.
A second reason to think that anti-corruption promises are an effective vote-winning strategy is that candidates with an anti-corruption agenda have recently been electorally successful. Promises to crack down on corruption are a central component in the anti-elitist rhetoric espoused by populist candidates. Anti-elite rhetoric is a form of political discourse that adopts a critical stance against individuals and groups perceived to be part of the political establishment. It typically involves portraying elites as corrupt, self-serving, or disconnected from the interests and concerns of the ordinary citizen (Rooduijn Reference Rooduijn2018). This hostility can be directed towards elected politicians, established parties, government officials, or large corporations (Curini Reference Curini2020; Inglehart and Norris Reference Inglehart and Norris2016).
The aim of anti-elite rhetoric is to rally public support by appealing to the frustrations, grievances, and distrust that people may have toward those in positions of authority or privilege. Populist candidates have successfully won elections by demonizing the political establishment and accusing their political opponents of rigging the system. Donald Trump’s 2016 U.S. presidential campaign politicized corruption by promising to “drain the swamp” and claiming that his Democrat opponent Hillary Clinton “must be the most corrupt person ever to run for president” (Flitter Reference Flitter2016). He used false accusations to portrait himself as the savior who is going to fix a “rigged” system (Gambino Reference Gambino2016). Similarly, the left populist candidate in the 2018 Mexican presidential election, Andrés Manuel López Obrador (AMLO), made the fight against corruption the core of his electoral campaign. He reduced the solutions to all of the country’s economic, social and political problems to the simple idea of fighting corruption (Méndez et al. Reference Méndez, Lendo and Flores2021). In sum, successful populist candidates, both right and left-leaning, have employed the promises to fight corruption as way to challenge the status quo and position themselves as champions of the ordinary folk.
While the anti-corruption rhetoric is an essential part of the populist discourse against the elite, it is not circumscribed to the populist framework, nor populism can be reduced to an anti-elite rhetoric. The anti-establishment element in the populist discourse, whereby the elite is undermining the popular will for selfish purposes, cannot alone be equated to populism (Hawkins and Rovira Kaltwasser Reference Hawkins and Rovira Kaltwasser2017), and does not have be to espoused only by populist parties (Wiesehomeier Reference Wiesehomeier2017). While there is a connection between promises to combat corruption and populism, it is important to note that anti-corruption messages are not exclusive to populist leaders. Non-populist leaders can also utilize these messages by shifting the focus towards clean government reforms instead of merely accusing political opponents of corruption. Unlike populist discourse, anti-corruption messages do not necessitate the aggression of the populist leader and can refrain from idealizing the “virtuous and pure” people or relying on the notion of a “unified” popular will, which are often found within populist frameworks.
A third reason to suspect that anti-corruption promises can win votes is their widespread adoption by parties from various ideological backgrounds and political styles. Increasingly, a diverse array of parties, whether populist or not, are incorporating anti-corruption rhetoric into their platforms. Studies of expert surveys on party positioning find that anti-corruption pledges are embraced by parties and candidates across the political landscape (Polk et al., Reference Polk, Rovny, Bakker, Edwards, Hooghe, Jolly and Koedam2017). Many new parties focusing on fighting corruption have capitalized on the public’s distrust of established political parties (Bågenholm Reference Bågenholm2013). Even mainstream parties with moderate social and economic policies, which differ greatly from radical populist parties, campaign on political reform, transparency, and changing the way politics is conducted (Hanley & Sikk, Reference Hanley and Sikk2016). Hence, anti-corruption pledges are embraced by non-populist parties too, seeking to mobilize disillusioned citizens. This widespread usage indicates that parties perceive anti-corruption messages as a valuable electoral strategy.
In sum, campaigning on an anti-corruption platform can be an effective electoral strategy for three reasons. Voters are more likely to be swayed by promises that tackle problems about which they have intense feelings, and electoral messages centered around cleaning up the government align well with these concerns. Recent populist candidates have won elections by employing an anti-elite rhetoric that demonizes the political establishment and promises to combat corruption. Parties, whether populist or not, adopt anti-corruption platforms to tap into the frustrations and grievances of the general public. This discussion leads us to the first hypothesis:
H1 Anticorruption promises increase electoral support.
While the argument elaborated above provides compelling indication that anti-corruption promises can win votes, it is essential to acknowledge that this outcome is not guaranteed. Anti-corruption promises might not necessarily lead to an increase in electoral support for several reasons. First, electoral promises can be ineffective if they are perceived as unreliable commitments. Voters are more likely to be swayed by specific proposals that outline concrete policies. Second, the effectiveness of anti-corruption platforms can also be undermined by a history of unfulfilled promises and voter skepticism. When citizens are deeply distrustful of parties’ electoral promises, partly because they have witnessed several policy switches by candidates who do not follow through on campaign pledges (Campello Reference Campello2014; Stokes Reference Stokes2001), anti-corruption promises may lack persuasive appeal. This could be particularly expected in countries where corruption is prevalent. Therefore, it remains uncertain whether anti-corruption promises will consistently translate into electoral gains.
4. When is Anticorruption Most Persuasive?
If anti-corruption messages during the campaign are not always an effective electoral strategy to secure votes, it becomes crucial to understand the circumstances under which voters are most likely to be swayed by anti-corruption appeals. The following two sections investigate key factors that can influence the persuasiveness of anti-corruption promises, namely the content of the message and the crebility of the messenger. In this section, I argue that certain characteristics related to the content of the message can make some anti-corruption platforms more compelling than others. Promises that are specific in outlining the actions, policies, or reforms to combat corruption tend to resonate more with voters. Vague or general promises, on the other hand, may be viewed as lacking substance and may not have the same persuasive impact.
For anti-corruption platforms to be compelling pledges, they must go beyond mere rhetoric. Voters are more likely to be persuaded if the candidate outlines concrete measures and policies that demonstrate a clear and strategic vision to cleaning up the government. However, electoral promises are often unsubstantiated statements devoid of policy content. Campaign promises can sometimes be considered a mere act of promising, or pledging, to address a particular issue (Aragones and Palfrey Reference Aragones and Palfrey2004). Moreover, voters suspect that candidates are ready to make whatever statement they believe will increase their probability of election, regardless of what they intended to do if elected. If electoral platforms constitute unreliable commitments, voter’s beliefs about what a candidate will do if elected might be unaffected by campaign promises. Hence, campaign promises need to be credible if candidates expect electoral messages to gain them votes.
One of the fundamental problems politicians face in the electoral arena is to find ways to make credible their appeals to voters. Given that parties are unable to make biding promises in the absence of external enforcers, candidates and parties might renege on their promises (Becher Reference Becher2016; Iversen and Soskice Reference Iversen and Soskice2006). One costly way to make electoral promises more credible is to develop a policy plan that could later serve voters as a yardstick to evaluate their reelection.
Endorsing a particular policy on a given issue thus sends voters the signal a candidate is serious about his intentions and has the necessary expertise to address such issue. By taking a firm stance and outlining concrete policy proposals, candidates can convey their competence and credibility to the electorate. When a candidate outlines a specific policy, it demonstrates that she has thoughtfully considered the problem and is willing and capable to take decisive action to tackle it. These anti-corruption measures could include strengthening regulatory frameworks, enhancing transparency and accountability, implementing anti-corruption reforms, and promoting integrity in public administration.
Moreover, endorsing a concrete policy on a given issue can help candidates differentiate themselves from their opponents. It allows them to articulate their unique approach and showcase their expertise or understanding of the matter. Importantly, when candidates endorse specific policies, it enables accountability. Voters can hold them responsible for delivering on their promises once elected. It establishes a benchmark against which the candidate’s performance and progress can be evaluated, allowing voters to assess whether the candidate has followed through on their commitments.
In contrast, when candidates resort to vague promises without providing substantial policy content, it sends a signal to voters that the candidate may not be capable of fulfilling their electoral pledges. The absence of specific details and concrete plans can undermine the credibility and trustworthiness of the candidate’s campaign. Although populist politicians often play the anti-corruption card to make a discursive point only, relying on the symbolism of fighting a corrupt elite, vague promises can actually leave voters uncertain about the candidate’s true intentions, competence, and the feasibility of their proposed actions. This lack of clarity can lead to skepticism and a perception that the candidate is avoiding accountability for their promises. This discussion leads to the following hypothesis:
H2 Anticorruption promises with policy content increase the candidate’s perceived effectiveness in reducing corruption more than anticorruption promises without policy content.
This study focuses on transparency policies as an example of a concrete anti-corruption promise with policy content. Not only is transparency widely regarded as an effective tool for reducing corruption and enhancing democratic quality (Florini Reference Florini2007; Hood and Heald Reference Hood and Heald2006), but the global adoption of FOI laws reflects a consensus between international organizations and domestic civil society on the importance of promoting transparency. These laws carry symbolic power, signaling politicians’ commitment to transparency and introducing a legal framework that is difficult to dismantle (Berliner Reference Berliner2014). Transparency policies, moreover, can empower civil society and the media, creating advocates for their continued implementation and advancement. Another reason for adopting transparency as the focus of anti-corruption promises in this study is the extensive body of research examining institutional, political, and cultural factors that drive greater transparency (Berliner Reference Berliner2014; Michener Reference Michener2011, Reference Michener2015; Schnell and Jo Reference Schnell and Jo2019; Williams Reference Williams2009). For instance, Berliner (Reference Berliner2014) argues that governments facing political competition and uncertain tenures are more likely to institutionalize transparency to secure access to information after leaving office. While existing research highlights the incumbent’s political incentives for implementing transparency, this study investigates when and why anticorruption pledges like transparency resonate with voters. Understanding public support for these pledges at the electoral stage can shed light on how social demand pressures politicians to adopt and sustain such reforms once in power.
5. Credible Anticorruption Messengers
Who the candidate is can also influence the persuasiveness and effectiveness of anti-corruption pledges. In this section, I argue that certain characteristics of the messenger can make some anti-corruption platforms more compelling than others. This theoretical expectation hinges on the notion that the credibility of the candidate is an influential factor, determining whether anti-corruption pledges will effectively sway voters. This section discusses the role of two candidate characteristics that can influence the persuasiveness and effectiveness of anti-corruption pledges: having a clean track-record and being a woman. I postulate that candidates with no history of involvement in corruption scandals are more credible anti-corruption candidates than politicians involved in corruption scandals. Similarly, I propose that female candidates will be a credible messenger of anti-corruption promises, as prevailing female stereotypes dictate that women are a cleaning force in government.
Electoral promises to clean up government can be more compelling if they come from candidates with no history of involvement in corruption scandals. Candidates with a record of personal integrity or a history of not being involved in any corruption scandals are seen as more trustworthy and capable of delivering on their promises to fight corruption. In contrast, anti-corruption promises can be overshadowed by corruption accusations. Candidates with questionable backgrounds or suspected involvement in corrupt activities may face heightened scrutiny and skepticism, undermining the impact of their anti-corruption messaging. Research shows that parties deeply involved in corruption scandals are not seen as the most trustworthy when it comes to fighting corruption (Bågenholm and Charron Reference Bågenholm and Charron2014).
It is not uncommon for self-proclaimed anticorruption candidates to be themselves involved in corruption investigations. The right-wing populist leader, Jair Bolsonaro, is a good example. He won the 2018 presidential election in Brazil on an anti-corruption platform, but he faced corruption accusations during his term (Milhorance Reference Milhorance2022; Phillips Reference Phillips2022). These mediatic episodes create reputational problems that weaken a candidate’s ability to claim a policy priority on an issue. After a scandal, the candidate would not be perceived to be the most competent at handling this particular problem or delivering the desired outcomes. Hence, voters are more likely to swayed by anti-corruption messages coming from candidates with a reputation of integrity or a clean record.
H3 Anticorruption promises by clean candidates increase electoral support and the candidate’s perceived effectiveness in reducing corruption more than anticorruption promises by corrupt candidates.
In addition, electoral promises to crack-down on graft and government improprieties can be more persuasive and effective if they come from female candidates. Gender stereotypes lead people to view women as a cleaning force in government and more dedicated to honest government (McDermott, Reference McDermott1998). Research shows that voters concerned with the levels of honesty and integrity in government have higher tendency to vote for women candidates (Dolan Reference Dolan2004). According to Stensöta and Wängnerud (Reference Stensöta and Wängnerud2018), in fact, the documented correlation between more women in government and lower perceived corruption led some policymakers to conclude women are better at combating corruption than men. Studies have shown, that female politicians mitigate the concerns about corruption because gender stereotypes dictate that they are perceived to be more ethical, honest, and trustworthy (Barnes and Beaulieu Reference Barnes and Beaulieu2014). As women are viewed as “cleaner” than men, governments often use women as anti-corruption signals (Armstrong et al. Reference Armstrong, Barnes, O’Brien and Taylor-Robinson2022).
Several factors might motivate these gender stereotypes and the corresponding anti-corruption strategies. First, women politicians are held to higher standards at the ballot box and thus exhibit more risk aversion to illegal activities than men (Esarey and Schwindt-Bayer Reference Esarey and Schwindt-Bayer2018). Second, women are excluded from the male-dominated networks of clientelism and hence are seen as a potential for disrupting these networks (Stensöta and Wängnerud Reference Stensöta and Wängnerud2018). Regardless of their origins, these preconceptions about how women in politics behave, or ought to behave, lead voters to view female candidates as being more effective in cracking down on corruption.
H4 Compared to male candidates, anticorruption promises by female candidates increase their electoral support and their perceived effectiveness in reducing corruption.
It is worth noting that, while voters might use gender as a shortcut to infer traits, beliefs, and issue positions of candidates, existing research provides limited guidance on how prevalent gender stereotypes are and how they relate to anti-corruption agendas. It could be argued that a candidate’s gender might be irrelevant or even detrimental when voters assess their ability to implement anti-corruption policies. According to K. Dolan (Reference Dolan2014), gender may not play a central role in candidate evaluations or voting decisions. Conversely, stereotypes that suggest men are more assertive leaders could disadvantage female candidates, portraying male candidates as better suited for challenging roles such as combating crime and corruption (Huddy and Terkildsen Reference Huddy and Terkildsen1993; Sanbonmatsu Reference Sanbonmatsu2002). If gender stereotypes work in this way, female candidates might not gain any electoral advantage, or might even be at a disadvantage when promoting anti-corruption platforms.
6. Empirical Approach
a) The case of Paraguay
Paraguay provides a particularly instructive case to examine whether and how anti-corruption promises work. First, the widespread nature of corruption in Paraguay makes anti-corruption promises among electoral candidates a realistic event. On the 2022 Corruption Perceptions Index (CPI) reported by Transparency International, Paraguay scored 28 points on a scale from 0 (“highly corrupt”) to 100 (“very clean”), below the Latin American average of 41. Over the past decade, many candidates in local and national elections have emphasized anti-corruption and transparency, and the theme was prominently featured in the 2021 municipal elections.
Second, the consistent electoral support for the Colorado Party and the resiliency of the party machine makes Paraguay a difficult case for finding public support for anti-corruption measures. The Colorado Party has governed the country for most of the past eight decades relying on clientelism and patronage, and it won the 2023 presidential election once again. Despite high-profile corruption allegations, including U.S. sanctions against top figures like Horacio Cartes, the party retained power, winning the 2023 presidential election. Santiago Peña, the Colorado candidate, defeated an opposition that vowed to clean up government corruption. Similarly, in the 2021 municipal elections, a former mayor of Asunción, implicated in an scandal over irregular invoioces, won the election despite the controversy. The Colorado Party maintained territorial control in these elections, despite public discontent and the image of corruption associated with its candidates. Analysts argue that opposition parties, though active, were very fragmented and unable to present a united front (Duarte-Recalde & Gonzalez-Tiepermann, Reference Duarte-Recalde and Gonzalez-Tiepermann2021).
Finally, reducing corruption has remained challenging despite the implementation of anti-corruption measures, including transparency reforms. The FOI law adopted in 2014 was praised as a tool for accountability but lacked an independent oversight body and suffered from weak enforcement, particularly at the local level. While it has enabled journalistic investigations, these efforts have not translated into significant reductions in corruption. As a result, many Paraguayans view anti-corruption and transparency initiatives with skepticism. This cynicism, coupled with a resilient political machine, makes Paraguay a least likely case for anti-corruption effects. Understanding the factors that make anti-corruption promises effective in such a challenging environment can provide valuable insights for other Latin American countries.
b) Survey design
To evaluate how effective and persuasive are anti-corruption platforms, I conducted a survey experiment with 2,060 voting age Paraguayans (See Appendix A for details). The unforced conjoint experiment was embedded in an online survey administered by Offerwise, a professional market research firm with extensive experience in Latin America, in August-September 2021. Survey participants were invited by email and sampling quotas were established to ensure variance across gender, age, and education categories. In Appendix B, a comparison is made between the distribution of demographic variables in this sample and population-level variables for which comparable data is accessible. Even though the sample over-represents younger and more educated groups as commonly observed in online surveys, convenience samples remain valuable research instruments and can be used for making causal inferences (Mullinix et al. Reference Mullinix, Leeper, Druckman and Freese2015). Nevertheless, to address concerns regarding potential sample bias, I use representative matching to correct imbalances in the sample (Kuffuor et al. Reference Kuffuor, Visconti and Young2022). Appendix C demonstrates that this correction does not alter the key findings of the study.
c) Conjoint experiment
I employed an unforced-choice conjoint design to study citizen attitudes toward anti-corruption platforms across a range of candidate characteristics. This design allows us to study the multidimensional preferences underlying a citizen’s choice. By randomly varying multiple attributes of hypothetical candidates, we can causally examine the importance of the attributes of interest relative to other attributes (Hainmueller and Hopkins Reference Hainmueller and Hopkins2015). While survey experiments on vote choice might not directly mirror actual voting patterns in the real world (Boas et al. Reference Boas, Hidalgo and Melo2019; Incerti Reference Incerti2020), conjoint designs, by requiring respondents to choose between two candidates with multiple randomized attributes, may more closely simulate the trade-offs and choices voters face in actual elections (Hainmueller et al. Reference Hainmueller, Hopkins and Yamamoto2014). This design, therefore, represents a significant improvement over traditional vignette-based surveys, offering a more realistic approximation of voter behavior. For the purposes of this study, I randomly varied gender, anti-corruption platform, and disciplinary record. To make the hypothetical candidate more realistic, the conjoint design allowed me to also vary: age, education, party, ideology, and competence.Footnote 1 The levels for each attribute had equal probability of being randomly drawn and no restrictions were imposed. Since survey participants evaluated five pairs of candidates, I cluster the standard errors at the respondent level.
The main attribute of interest, the anti-corruption platform, had three levels. Individual respondents were randomly exposed to a baseline condition where the candidate’s party does not adopt any anti-corruption platform: “his/her party does not have anti-corruption proposals”,Footnote 2 or a condition where the candidate’s party adopts a vague anti-corruption platform: “his/her party proposes to fight corruption, but does not offer concrete anti-corruption measures,” or finally, a condition where the party does offer a specific anti-corruption policy: “his/her party proposes to fight corruption, and offers concrete anti-corruption measures such as making public administration more transparent.”Footnote 3 , Footnote 4 For the gender attribute of the hypothetical candidate, I used the binary: male/female.
Finally, the misconduct attributes involved two types of corruption, bribery and embezzlement. Following Ang’s typology (Ang Reference Ang2020a), where she distinguishes “access money” from “grand theft,”Footnote 5 I describe the candidate in the bribery treatment as having received bribes in exchange for public works contracts or having not received such bribes. The embezzlement attribute also consists of two levels. It portrays the mayoral candidate as having received the salary of “ghost” municipal officials by creating false positions or as having not misappropriated any money from the municipal government. See table A4 in the Supplementary Appendix for a complete list of all the attributes and levels.
After each pair of profiles, respondents answer two questions measuring a vote and an attitudinal outcome respectively. First, the vote outcome consists of a question asking about electoral support for the hypothetical candidate: “If the general municipal elections were held tomorrow, which of these candidates would you vote for?” Respondents are given the option to choose one of three alternatives: Candidate 1, Candidate 2, or would not vote. Recognizing that vote choice can be a demanding or strenuous decision, I opted not to force respondents to choose one candidate. Recent methodological studies recommend using unforced choice designs to allow respondents to opt out if they lack an opinion or disapprove of all presented options (Miller and Ziegler Reference Miller and Ziegler2024; Visconti and Yang Reference Visconti and Yang2024). This approach acknowledges the realistic alternatives of abstaining or casting a blank/null vote, helping avoid misclassification errors and improve external validity (Visconti and Yang Reference Visconti and Yang2024). Notably, empirical research increasingly incorporates unforced choice outcomes in conjoint experiments to study preferences for candidates, policies, or migration locations (Agerberg Reference Agerberg2020; Eggers et al. Reference Eggers, Vivyan and Wagner2018; Frederiksen Reference Frederiksen2024; Ghosn et al. Reference Ghosn, Chu, Simon, Braithwaite, Frith and Jandali2021; Howard and Wehde Reference Howard and Wehde2023). For this outcome, I coded absentee votes as missing values following Visconti and Yang (Reference Visconti and Yang2024), to reflect the empirical reality that the preferences of non voters are inherently unobservable. As shown in Appendix E the results are robust to coding no-vote responses as zeros, as done in other studies (Agerberg Reference Agerberg2020; Miller and Ziegler Reference Miller and Ziegler2024; Frederiksen Reference Frederiksen2024). Although treating abstentions as missing entails some loss of information, the primary aim of this study was not to analyze abstention or the decision to refrain from voting. Rather, the focus was on presenting respondents with a realistic choice set and, by excluding non-votes, aligning the sample more closely with the target voting electorate. Future research should leverage opt-out responses to better understand the conditions under which respondents choose not to select any of the available options.
The attitudinal outcome in this study aims to assess the perceived candidate’s effectiveness in reducing corruption. To evaluate the factors influencing the perceived effectiveness, participants were presented with the following question: “Which of these candidates do you believe will help reducing corruption the most?” Respondents had to choose between candidate 1 and candidate 2. This question seeks to capture respondents’ belief in a candidate’s ability to implement promised anti-corruption measures and achieve the desired outcome of reducing corruption.
To analyze the conjoint experiment, I estimated the average marginal component effect (AMCE) and the marginal means (MMs). The AMCE is a causal estimand that represents the marginal effect of a specific attribute over the joint distribution of all the other attributes (Hainmueller and Hopkins Reference Hainmueller and Hopkins2014). For example, the AMCE for the attribute value of a candidate’s concrete anti-corruption platform measures the average difference in the probability of a candidate winning electoral support if he/his adopts a concrete anti-corruption platform compared to if he/she does not adopt one (baseline), where the average is computed given all other possible combinations of the other attributes. The MMs, in turn, give the probability that respondents chose a candidate with a given attribute level, and are particularly useful for subgroup analysis (Leeper et al. Reference Leeper, Hobolt and Tilley2020).
7. Main Results
A first step to examine the effectiveness of anti-corruption platforms involves looking at whether, all else equal, anti-corruption candidates are preferred to candidates who do not adopt any anti-corruption platform (baseline). Figure 1 displays the effects of the concrete and vague anti-corruption platforms on the vote outcome: electoral support (Hypothesis 1). The dots indicate point estimates and the lines represent 95% confidence intervals for the AMCE of each attribute value on the probability the respondents chose a candidate for mayor. The dots without the confidence intervals denote reference categories.
Effects of Anticorruption Platform on Probability of Electoral Support.
Note: This figure displays the AMCEs for the attribute of interest. See Supplementary Appendix for full model summary in table format.
When selecting candidates for mayor, the results demonstrate that candidates who advance a concrete anti-corruption platform are 12 percentage points more likely to win electoral support than candidates who do not pledge to fight corruption. Vague anti-corruption promises also earn the hypothetical candidate an electoral advantage; these candidates are 4.5 percentage points more likely to win electoral support than candidates who do not present any anti-corruption platform. This evidence provides support for H1, according to which promising anti-corruption measures is, on average, an effective electoral strategy for winning votes.
In relation to other attributes, the effect of anti-corruption promises on electoral support, whether vague or concrete, is smaller than having a bribery record (17 percentage points), or an embezzlement record (19 percentage points). This confirms the strong negative effect of corruption demonstrated in previous survey studies, none of which focused on anti-corruption promises (Avenburg Reference Avenburg2019; Breitenstein Reference Breitenstein2019; Klašnja et al. Reference Klašnja, Lupu and Tucker2021; Vera Reference Vera2020; Weitz-Shapiro and Winters Reference Weitz-Shapiro and Winters2017). Moreover, the effect of an anti-corruption platform is larger than the effect of ideological closeness (2 percentage points) or gender (4 percentage points). The surprisingly small effect of ideology is likely due to the absence of specific party names in the design; however, the large effect of anticorruption promises is still substantial and should not be overlooked.
It is worth noting that the rate of no-vote responses was 33% which resembles the abstention rate of 40% observed in Paraguay’s 2021 municipal elections (COPPPAL, 2021). This suggests that including a no-vote option yielded a realistic pattern of abstention. While not the focus of this study, a preliminary analysis (Appendix M) shows that abstention is more likely when a candidate is associated with bribery and embezzlement, and less likely when a candidate presents a concrete anticorruption platform, is competent and has college education. Among respondent characteristics, political ideology is significantly associated with the likelihood of abstaining, with conservative respondents less likely to opt out of voting. Future research could build on this underexplored aspect of the unforced-choice conjoint design to better understand the factors that lead respondents to opt out of selecting any of the available alternatives.
A second step to understanding the effectiveness of anti-corruption promises is looking at the second outcome: which candidate citizens believe will help reducing corruption the most. I argued that anti-corruption promises would be more persuasive if they can convince voters that the candidate intends to and is capable of successfully implementing anti-corruption policies if elected. Since endorsing a particular policy on a given issue sends voters the signal a candidate is serious about his intentions and has the necessary expertise, anti-corruption platforms with a clear policy stand will be more likely to increase the candidate’s perceived effectiveness in reducing corruption than platforms without specific anti-corruption proposals (hypothesis 2).
Figure 2 shows that anti-corruption promises, both vague and concrete, have a positive effect on the perceived candidate’s effectiveness in reducing corruption. Candidates that adopt a concrete anti-corruption platform are 13 percentage points more likely to be perceived as a candidate that will successfully reduce corruption than those that do not leverage such promises. As expected, concrete anti-corruption platforms are more persuasive of the candidate’s pledge to fighting corruption than vague proposals. Meanwhile, vague anti-corruption promises only increase the perceived effectiveness by 7 percentage points. As expected, the effect of anti-corruption promises with policy content on the belief the candidate will reduce corruption is greater —almost twice as large— as promises without policy content.
Effects of Anticorruption on Perceived Effectiveness in Reducing Corruption.
Note: This figure displays the AMCEs for the attribute of interest. See Supplementary Appendix for full model summary in table format.
This evidence provides support for H2. When candidates rely on ambiguous promises without offering substantial policy details, it conveys a message to voters that the candidate might lack genuine intentions or the necessary expertise to fulfill their electoral commitments. This suggests that respondents do not respond as favorably to vague anti-corruption promises as they do to concrete proposals to increase transparency in government.
It is worth noting that support for concrete anticorruption platforms is stronger among respondents dissatisfied with the political establishment (see Supplementary Figures A21 and A22). This heterogeneous treatment effect aligns with the notion that anticorruption pledges are electorally effective in settings where corruption is a salient issue and a broad range of parties use anticorruption promises to appeal to disillusioned citizens.
8. Who is a Credible Anticorruption Messenger?
I now turn to the analysis of whether the personal traits of the candidate can make some anti-corruption promises more compelling than others. Figure 3 shows the effect of anti-corruption platforms on a candidate’s perceived effectiveness in reducing corruption by candidate misconduct. The left-hand panel depicts the AMCEs, and the right-hand panel reports the MMs. While AMCEs in subgroup analysis allow us to establish differences in effect size, MMs can help us visualize differences in preferences or levels of support (Leeper et al. Reference Leeper, Hobolt and Tilley2020).
Effect of Anticorruption Platform on Perceived Effectiveness in Reducing Corruption by Bribery Record.
I find that, contrary to hypothesized, the effect of anticorruption on perceived candidate’s effectiveness in reducing corruption does not vary depending on their bribery record. Interestingly, the anti-corruption platform is persuasive of the candidate’s anti-corruption effectiveness, regardless of a candidate’s misconduct. Among candidates with a clean record, the anticorruption platforms increase a candidate’s perceived anticorruption effectiveness by 7.9 and 13.9 percentage points. The effect sizes are similar for candidates with a corrupt record, as seen by the overlapping confidence intervals (6.9 and 12.4 percentage points). Both candidates, those with a history of taking bribes and those without it, who endorse an anti-corruption platform increase voter’s perception that the candidate will fulfill its commitment.Footnote 6 However, this does not mean that corrupt candidates are preferred over clean ones. The MMs in the right-hand panel reveal that the perceived candidate’s effectiveness in reducing corruption among corrupt candidates is significantly lower (35.7, 42.5 and 48.2 percentage points) than among clean candidates (50.5, 58.5 an 64.4 percentage points). The p-values associated with the differences in marginal means within each anticorruption level are p=0.000. While citizens do not significantly credit clean candidates for their anticorruption stances more than they do corrupt candidates (per AMCE), citizens do prefer clean anticorruption candidates over corrupt anticorruption candidates (per MM).
Next, when examining the vote outcome, electoral support, I observe comparable findings. The left panel of Figure 4 shows that there is no significant difference on electoral support for the anti-corruption platform across candidates with different misconduct records. Both candidate types, with and without bribery record, are rewarded electorally for anti-corruption proposals.Footnote 7 The right panel, shows the subgroup levels of support. As expected, candidates with a clean record are more supported than candidates with a bribery record. For example, a corrupt candidate who adopts a concrete anticorruption platform is preferred by 47 percentage points, while a candidate with a clean record who does not adopt such platform is preferred by 65 percentage points. Similarly, a candidate with a vague anticorruption platform exhibits greater support if he has a clean record than if he has a corrupt record (57 versus 39 percentage points). This means that while anticorruption advantages both corrupt and clean candidates, the overall levels of support for clean candidates are still greater than for corrupt ones. In other words, corrupt candidates receive an electoral boost for adopting an anticorruption platform, but this boost does not make their electoral favorability exceed that of clean candidates.
Effect of Anticorruption Platform on Electoral Support by Bribery Record.
Overall, the evidence provided in Figures 3 and 4 does not provide support for H3 that candidates with a clean record receive a greater anticorruption reward than politicians with a bribery record. I failed to observe any significant differences in the effect of anticorruption for bribe-taking and clean candidates. Instead, a more nuanced picture of anticorruption dynamics emerges. Having a clean record does not make anticorruption pledges more persuasive, but clean candidates, regardless of their anticorruption stances, are still preferred over corrupt ones.
Finally, hypothesis H4 posited that gender could play a role in how persuasive and effective anti-corruption promises are. The left panel of Figure 5 shows the effects of an anti-corruption platform on perceived candidate’s effectiveness in reducing corruption by the gender of the candidate. Surprisingly, I do not find evidence that women candidates are more credible anti-corruption messengers than male candidates. In fact, anti-corruption messages are slightly more persuasive when delivered by men, though the difference is not statistically significant. The right panel of Figure 5 shows the marginal means and confirms that women’s mean level of perceived effectiveness is not significantly higher than men’s, except at baseline. When both candidates lack an anticorruption agenda, respondents view women as more effective at reducing corruption. However, when candidates adopt an anticorruption platform, men’s perceived effectiveness matches that of women. This result is almost the opposite of what was hypothesized, as men manage to overcome their initial disadvantage when they promise to combat corruption.
Effects of Anticorruption Platform on Perceived Effectiveness in Reducing Corruption by Candidate Gender.
While the above results suggest that anticorruption platforms coming from female candidates are equally persuasive, if not less so, than those coming from male candidates, do anticorruption female candidates nonetheless garner more electoral support? Figure 6 presents the effect of an anticorruption agenda on electoral support by candidate gender. The left panel shows the conditional AMCEs for men and women, and the right panel shows the marginal means for men and women candidates. The AMCEs reveals that vague anticorruption promises by male candidates are rewarded electorally but the reward for vague anticorruption promises by female candidates does not reach statistical significance. The marginal means provide a clearer view of this dynamic. Male candidates who adopt an anticorruption discourse gain an electoral advantage that female candidates do not. For instance, men with a concrete anticorruption platform exhibit similar levels of electoral favorability as women (58 for women and 55 for men). Likewise, the favorability of women with a vague anticorruption platform is nearly the same as that of men (49.6 for women and 47.5 for men). Even if women were initially preferred over men by a wide margin at baseline, this advantage disappears as soon as candidates adopt anticorruption platforms. This means that while anticorruption messages can gain a male candidate substantial electoral support, these promises are less effective vote-winning tools for female candidates. This echoes recent experimental findings that citizens are more responsive to performance information about male leaders (Diaz et al. Reference Diaz, Oliveros, Weitz-Shapiro and Winters2023).
Effects of Anticorruption Platform on Electoral Support by Candidate Gender.
Overall, when considering the results from Figures 5 and 6, it becomes clear that the evidence does not support the hypothesis that women are more credible anticorruption advocates than men. The perceived anticorruption effectiveness of women does not receive a greater boost than that of men when adopting anticorruption platforms. In fact, women appear to lose their initial electoral advantage; when candidates adopt anticorruption platforms, men reach the same level of favorability as women. This vanishing of women’s advantage suggests that anticorruption promises are less effective vote-winning tools for female candidates. Therefore, there is some indication that only men benefit electorally from anticorruption platforms, which suggests a gender bias against women, not in favor of them.
9. Conclusion
This paper argues that anticorruption messages are an important component of voters’ decision-making and can significantly shape electoral accountability. The empirical findings indicate that adopting an anticorruption rhetoric can secure votes for candidates, even though promises to reduce corruption may appear unfeasable or unreliable to an skeptical electorate. Rather than dismissing anticorruption appeals, a candidate’s firm stance against corruption resonates with voters’ dislike of corruption. Importantly, in terms of persuasiveness, voters find concrete anticorruption commitments more compelling than empty rhetoric. This indicates that electoral accountability may be achievable, as politicians have incentives to develop their platforms with policy content, thereby providing citizens with the tools to assess anticorruption candidates more effectively in the future.
While the policy incentives faced by candidates might initially suggest a promising outlook for democratic accountability, the finding that both clean and corrupt candidates are similarly rewarded for anticorruption rhetoric is troubling. This suggests that corrupt candidates can ironically mitigate the repercussions of their corruption by adopting anticorruption discourse. It may also suggest that in highly corrupt environments like Paraguay, citizens might perceive corruption as a normal practice, unworthy of being combated or denounced (Corbacho et al. Reference Corbacho, Gingerich, Oliveros and Ruiz-Vega2016; Persson et al. Reference Persson, Rothstein and Teorell2013). If most citizens view corruption as a social norm (Köbis et al. Reference Köbis, Iragorri-Carter and Starke2018, Reference Köbis, Troost, Brandt and Soraperra2022), they may be less inclined to take action against corrupt politicians, treating them no differently than those who are not corrupt.
Additionally, the results indicate that only male candidates benefit from anticorruption rhetoric. Anticorruption promises coming from women are just as convincing of a candidate’s effectiveness as those men, if not less so. Even if women have an electoral advantage when no candidates promise to reduce corruption, this advantage disappears when candidates adopt anticorruption platforms. Therefore, anticorruption platforms can be considered less effective tools for female candidates, since they offer male candidates a more substantial electoral boost.
These findings have important implications for the research on political corruption and voting behavior, as they draw attention to an overlooked aspect of the electoral accountability for corruption: the impact of anticorruption promises. Given the salience of corruption as an electoral issue, we should examine not only candidates implicated by corruption accusations but also those who propose remedies to systemic corruption. This paper addresses this research gap by systematically analyzing the appeal of anticorruption platforms. It explores how anticorruption discourse can be appropriated by populist leaders, likely due to its alignment with anti-elite sentiments, highlighting the risks of corruption’s extreme politicization. Crucially, the study reveals a concerning pattern: anticorruption platforms serve as equally effective and persuasive tools for both clean and corrupt candidates. While clean candidates are favored due to their higher baseline favorability, both receive a similar boost from such platforms. This offers a nuanced answer to whether anticorruption platforms allow corrupt politicians to avoid accountability. Although anticorruption campaigns do not fully reverse opposition to corrupt candidates—voters still prefer clean ones with credible anticorruption agendas—they can significantly help corrupt politicians reduce, though not eliminate, electoral sanctions. This is a key contribution to the literature on factors mitigating accountability for corrupt politicians.
Moreover, the findings suggest that anticorruption reforms are likely to encounter minimal public resistance if genuinely committed politicians implement concrete policy initiatives. Given the importance of effectively combating corruption, this paper underscores the critical need to understand public support for anticorruption measures. Notably, this study demonstrates that there is a social demand for government transparency, even in a context marked by skepticism and cynicism. However, it remains uncertain whether citizens’ support for transparency pledges in this experiment extends to other anti-corruption measures. Future research should explore public approval of various types of anti-corruption policies.
For researchers studying populism, this study provides compelling evidence that a specific aspect of the populist discourse can wield persuasive power on its own, without resorting to nasty attacks against the opponent or idealizing the “virtuous” people. In fact, respondents were more persuaded of a candidate’s commitment to clean the government by concrete anticorruption pledges offering specific reforms than by vague anticorruption pledges. This might seem surprising given the numerous examples of populist leaders winning elections with vague anti-corruption promises. However, it is likely that the lower visibility and media coverage of candidates with concrete anticorruption plans creates the impression that these candidates are less effective than their populist counterparts. Nevertheless, the present research highlights the value of isolating a specific aspect of the populist discourse, shedding light on the features that resonate most with voters. While previous research has focused on the demand side of populism (Akkerman et al. Reference Akkerman, Mudde and Zaslove2014; Bakker et al. Reference Bakker, Schumacher and Rooduijn2021; Castanho Silva et al. Reference Castanho Silva, Jungkunz, Helbling and Littvay2019; Hawkins et al. Reference Hawkins, Kaltwasser and Andreadis2020), this study delves into the persuasive qualities of anticorruption messages. This emphasis is in line with recent research on anti-establishment messaging (Hansen and Treul Reference Hansen and Treul2021).
For researchers of gender and corruption, this study offers new insights into how female politicians are disadvantaged in electoral competition, specifically regarding anticorruption pledges. The finding that anticorruption pledges disappear women’s electoral advantage challenges the prevailing view in the literature that women are viewed more favorably on integrity issues. Not only does the perceived effectiveness of anticorruption efforts fail to increase more for women than for men, but a slight gender difference in vote outcomes also emerges, favoring men. It is possible that gender stereotypes portraying men as more assertive leaders undermine the advantage women typically have on integrity issues. Additionally, we cannot rule out the possibility that citizens may be more responsive to campaign messages from men than from women. Regardless, if male candidates benefit more from anticorruption stances than female candidates, it raises the urgent question of what role women can play in reforming corruption and what their prospects for doing so are (Barnes and Beaulieu Reference Barnes and Beaulieu2024).
There are, of course, many limitations to this study. First, this paper provides only two versions of an anticorruption platform, deliberately avoiding the incendiary language commonly associated with such pronouncements. As such, the study does not explore the broader electoral impact of other components typically found within the populist package. This study focuses solely on the stoic and emotionless aspect of the anticorruption appeal, isolating the “clean-up government” message to avoid the divisive language often used by populist leaders. Future studies should investigate the other components of the populist discourse for a more comprehensive understanding.
Second, the study offers only a brief exploration of the characteristics of citizens most receptive to anticorruption promises. The findings suggest that disillusioned and disenchanted citizens are more likely to support anticorruption platforms. However, future research should delve deeper into this area to identify the specific demographic, psychological, or social traits that make certain voters more susceptible to anticorruption appeals. Understanding who is most influenced by these messages is crucial for gaining a deeper understanding of how these platforms shape electoral outcomes.
Finally, future investigations should explore the external validity of the findings presented here. It is possible that Paraguayans are more skeptical of anticorruption promises due to the country’s high level of corruption. In highly corrupt settings, citizens tend to be cynical, doubting that any party can competently address corruption (Pavão 2018). In scenarios with lower corruption, in contrast, citizens might be more trusting of anticorruption promises. It is possible that in less corrupt countries citizens are more susceptible to being swayed by anticorruption pledges to clean up government, as they have not witnessed past promises failing to materialize. Further research in diverse settings is essential to gain a comprehensive understanding of how anticorruption appeals work across political contexts.
Supplementary material
Supplementary material for this article can be found at https://doi.org/10.1017/S0007123425100501
Data availability statement
Replication data for this article can be found in Harvard Dataverse at https://doi.org/10.7910/DVN/66ANEK
Acknowledgements
The author gratefully acknowledges Miguel Carreras, Gustavo Diaz, Kevin Mullinix, Giancarlo Visconti, and participants at the 2024 MPSA panel for their valuable feedback on the experimental design and earlier drafts of this paper.
Finacial support
This work was supported by a grant provided by the National Council of Science and Technology of Paraguay (CONACYT).
Competing interests
The author declares no competing interests.
Open access
