The Fairer Sex vs. the Fairer System: Improving the Narrative Between Women’s Increased Political Participation and Reduced Corruption

1. Executive Summary

Transparency International’s 2024 index report revealed that corruption remains a significant threat worldwide. As this struggle continues, assumptions around women’s increasing level of political participation arise, notably with the claim that ‘women are less corrupt than men’. Repeated studies such as Dollar, Fishman and Gatti (2001) and Swamy, Knack, Lee, and Azfar (2001) have found that higher female political representation does indeed correlate with lower corruption. This finding has widely been cited by international organizations and anti-corruption advocates such as Transparency International and the United Nations Office on Drugs and Crime (UNODC), where increased political participation is used as a measure against corruption. However, this correlation does not directly imply an automatic causation. This report investigates the claim that there is a ‘fairer sex’, rather than just assuming it. Through statistical, experimental, and case-based data, this analysis argues that institutions play a much greater role in corruption than gender. The relationship between women’s political participation and decreasing levels of corruption is not tied to the latter, but rather shaped by the political environment that enables it.

The Issue Synopsis establishes three principles: First, that the correlation between women’s political representation and reduced corruption is genuine and has been replicated across multiple methods, from cross-country regressions to controlled laboratory bribery experiments. The two studies that popularized this correlation moved from individual attitudes to aggregate policy claims without actually demonstrating the mechanism connecting the two. An “individualist fallacy,” this report takes seriously rather than repeats. Second, once the relevant institutions are properly accounted for, and according to Sung’s (2003) national cross-analysis, it disappears into statistical nonsignificance altogether. Third, and most importantly, it is not democratic accountability in general that has an impact, but specifically judicial independence and press freedom. Electoral competition, on its own, is weakly associated with more corruption, not less. Contested elections without a watchdog judiciary and press simply create opportunities for vote-buying. Corruption itself frequently restricts women’s access to political office in the first place, meaning representation and corruption are entangled in both directions rather than following a one-way arrow.

The Insight and Analysis section explains why this happens, essentially building an ‘institutional mask’ over the gender effect. The appearance of a personal virtue that, upon closer inspection, is a product of the structure surrounding the person exercising it. This section argues that accountability institutions create the structures that make integrity politically survivable. Patronage-based political systems function through informal, male-dominated networks that structurally exclude women regardless of their individual ethics. Ignoring this risks reformers pouring their resources into representation quotas that generate symbolic progress without ever truly addressing corruption.

The Policy Recommendations section responds to this ‘diagnosis’ directly. Rather than treating gender-inclusion programs and anti-corruption programs as separate priorities, I recommend that international donors and domestic reformers sequence support for women’s political representation alongside targeted accountability-strengthening measures, especially in weak institutions. This policy reform aims to transform a symbolic initiative into a feasible approach to both include more women in politics and correctly target corruption.

2. Issue Synopsis

Introduction

It is important to begin to recognise that corruption is deeply rooted in patriarchal norms, generally making it harder for women to engage in corrupt practices. Historically, men have dominated positions of power, granting greater access to underground networks, patronage systems and informal connections, whereas women have fewer opportunities to engage in corruption. As a result, corruption stems less from an intrinsic moral difference among women than their exclusion from a ‘seat at the table’. Because of male-dominated governments, women have not yet had the full-fledged opportunity to carry out evident corruption, as observed in men.

As such, the idea that women’s political participation reduces corruption has settled within international development and governance circles. A closer look at the evidence, however, reveals a relationship that is real but conditional. Three thematic claims structure this analysis. First, the underlying correlation between women's representation and lower corruption is robust and has been replicated across different types of evidence. Second, the strength of this relationship depends heavily on whether democratic accountability institutions already exist around it. Third, corruption itself restricts women's access to political power, meaning representation and corruption are entangled in both directions rather than governed by a simple cause pointing one way. Taken together, these claims show why "increase women's political representation" cannot function as a standalone anti-corruption strategy on its own, and why institutional context has to be built into the policy from the start, not added as an afterthought.

Problem Analysis Framework

A. The correlation between women's political representation and lower corruption is real, and it has been replicated across both statistical and experimental research.

Cross-national statistical studies consistently find this relationship.

a) Dollar, Fisman, and Gatti's (2001) foundational cross-country study first established the correlation between women's participation in government. It reduced corruption and remains the most-cited reference point in the field to this day.

b) Esarey and Schwindt-Bayer (2019), using instrumental variable methods specifically designed to address concerns about causality, find a statistically significant reduction in corruption associated with increased women's representation.

c) Bauhr, Charron, and Wängnerud (2018) find that women's inclusion in local councils is strongly associated with reduced petty and grand corruption alike. d) The U4 Anti-Corruption Resource Centre's broader synthesis confirms this pattern holds across a wide range of national contexts, not merely a handful of convenient case studies.

Experimental and behavioral research shows the same pattern under controlled conditions.

a) Barnes and Beaulieu (2014) find that the mere presence of a female candidate reduces voters' likelihood of suspecting election fraud, independent of the actual circumstances of the election.

b) Alatas, Cameron, Chaudhuri, Erkal, and Gangadharan (2009) found Australian women measurably less tolerant of corrupt behavior than Australian men within laboratory bribery experiments.

c) Frank, Lambsdorff, and Boehm's (2011) review of laboratory corruption experiments finds that wherever a gender gap does appear, it consistently runs in one direction. No study in this literature finds men less corrupt than women.

d) Esarey and Chirillo (2013) find a measurable corruption-tolerance gender gap across a large cross-national attitudinal sample, lending further weight to the pattern.

It is worth noting that the two foundational studies establishing this correlation share a specific methodological weakness. Sung (2003) points out that Dollar, Fisman, and Gatti (2001) and Swamy et al. (2001) both move directly from an individual-level finding. Namely, that women as individuals tend to disapprove of bribery more than men do, to an aggregate-level policy claim. Namely, that seating more women in parliament will reduce national corruption, which he terms as an "individualist fallacy". In other words, an attitude held by individuals does not automatically scale up into an institutional outcome, and assuming that it does is the kind of cross-level leap that has misled policy debates in other fields as well.

B. The strength of this relationship is conditional on the presence of democratic accountability institutions, and once those institutions are properly accounted for, the correlation weakens considerably, or disappears altogether.

The gender gap in corruption tolerance is significantly larger in consolidated democracies than in autocracies.

a) Esarey and Chirillo (2013) find that the relationship between gender and corruption tolerance is strongest where corrupt behavior carries real electoral and judicial consequences.

b) The same study finds that this gap narrows considerably within autocratic systems, where corruption functions as a routine feature of governance rather than a punishable deviation from it.

c) Esarey and Chirillo interpret this pattern as evidence that behavior is shaped by institutional incentive rather than fixed disposition, a distinction this report treats as central.

d) Sung (2003) pushes this same logic considerably further, using cross-national regression data on 99 countries: once liberal democratic institutions enter the model, the female-representation-corruption relationship does not merely weaken; it becomes statistically nonsignificant across every specification he tests. Thus, the "fairer sex" effect and the underlying "fairer system" cause were never properly disentangled in the original studies to begin with.

Sung's (2003) findings go one step further still, breaking "liberal democracy" apart into its component institutions rather than treating it as a single undifferentiated variable.

a) Two components carry essentially all of the institutional weight in his analysis: rule of law and freedom of the press, both of which maintain strong, statistically significant negative relationships with corruption across every model he runs.

b) Electoral democracy, simply having contested elections, shows a positive, though not statistically significant, relationship with corruption in the same models.

c) Sung's own explanation is that elections alone create incentives for ‘vote-buying’ and ‘influence-peddling’, and only produce cleaner governance once paired with an independent judiciary and an investigative press capable of exposing wrongdoing after officials take office.

d) "Accountability" cannot be treated as one undifferentiated bundle. It is specifically judicial independence and press freedom, not electoral competition on its own, that appear to convert representation into reduced corruption.

Laboratory experiments reveal the same conditionality across cultural and institutional context.

a) Alatas et al. (2009) found that while Australian women were less tolerant of corruption than Australian men, no significant gender difference emerged among comparable samples in India, Indonesia, or Singapore.

b) The authors themselves conclude that previously reported gender differences may be culture and context-specific rather than universal.

c) The same study found greater cross-country variation in women's corruption-related behavior than in men's. This suggests that it is women's behavior, specifically, that responds most sensitively to institutional context.

d) Frank, Lambsdorff, and Boehm (2011) similarly conclude that observed gender effects depend on institutional and cultural setting, rather than reflecting some intrinsic gender trait carried across borders.

C. Corruption itself restricts women's access to political power in the first place, meaning representation and corruption are linked in both directions rather than by a simple, one-way cause.

Weak accountability, patronage-based systems actively limit women's entry into politics before representation ever becomes a question.

a) The U4 Anti-Corruption Resource Centre notes that male-dominated patronage networks hinder women's ability to run for office from the very start.

b) Bjarnegård (2013) finds that in clientelist systems, women are frequently excluded from corrupt informal exchanges specifically because they are not trusted as partners within them.

c) This exclusion reflects a structural barrier to entry rather than any form of personal integrity. It has nothing to do with individual ethics, and everything to do with who is allowed inside the room.

d) Bjarnegård's broader concept of "homosocial capital" argues that these informal networks are built, maintained, and reproduced by and for men already holding power, which is precisely why they are so difficult to break into from outside.

Statistical evidence supports a two-way causal relationship rather than a one-way effect running from representation to corruption.

a) Esarey and Schwindt-Bayer's (2019) instrumental variable analysis finds evidence of effects running in both directions: representation reducing corruption, and corruption independently reducing women's political participation.

b) This bidirectionality matters enormously in practice: increasing representation within a highly corrupt system may be harder to achieve in the first place, not simply less effective once achieved.

c) Bjarnegård also finds that informal networks can occasionally serve as a compensatory access route for women where formal institutions fail to guarantee equity, showing just how complex and non-uniform this relationship actually is.

d) Goetz's (2007) analysis of women as "political cleaners" reaches a similar conclusion, framing anti-corruption effects as dependent on women's structural position relative to existing power arrangements, rather than some innate quality carried into office.

III. Conclusion

Put together, these three claims establish that the widely cited link between women's representation and lower corruption is genuine, but conditional. The relationship is strongest, and perhaps only real, where specific institutions already exist: an independent judiciary and a free press, rather than electoral competition on its own. This raises the central puzzle to resolve: why do these particular institutions (not democracy in general, but rule of law and press freedom specifically) reshape individual behavior so thoroughly, and why do patronage networks specifically exclude women rather than simply failing to attract them? Understanding both of these mechanisms is necessary before any credible recommendation about gender-inclusion and anti-corruption policy can be made.

3. Insight & Analysis

Introduction

The Issue Synopsis established that women's political representation is associated with lower corruption, but only reliably so where democratic accountability already exists, and that corruption itself restricts women's access to power in weaker institutional settings to begin with. This raises the question: why does institutional context reshape a relationship so many assume to be a fixed matter of gender and disposition? This section examines the incentive structures accountability institutions create, the informal networks that structurally exclude women from corrupt exchange, and what happens when these two mechanisms collide or fail to meet at all.

Accountability Is Key

The starting error in most popular discussions of gender and corruption is treating corruption tolerance as though it were a fixed trait of character, rather than a behavior that bends according to consequence. Esarey and Chirillo (2013) find that the gender gap in corruption tolerance is considerably larger in democracies than in autocracies. A finding that is difficult to reconcile with any account resting on some essential, biologically rooted honesty. If women were simply more honest as a matter of disposition, that gap should hold steady no matter the regime type surrounding it. Instead, it nearly collapses wherever corruption goes largely unpunished. Essentially, the gap is not carried by the individual; it is carried by the system she operates within.

Sung's (2003) ‘reanalysis’ of the original "fairer sex" studies pushes this considerably further. Sung goes back to Dollar, Fishman, and Gatti (2001) and Swamy et al. (2001) and points out: both studies take an individual-level finding, that women personally disapprove of bribery more than men do, and use it to justify an aggregate-level policy claim, that electing more women will produce a less corrupt government. Sung calls this an ‘individualist fallacy’. An attitude held by individuals, in other words, does not automatically aggregate into an institutional effect. It has to be shown, not assumed.

Once Sung actually tests this using cross-national data from 99 countries, the assumed effect does not weaken; it disappears. When he introduces measures of liberal democracy into the same regression models used by the original studies, the statistical relationship between female representation and corruption drops out of significance entirely. But the more interesting move in his analysis is what happens when he refuses to treat "liberal democracy" as one undifferentiated set and instead breaks it into its component institutions. The rule of law and press freedom carry all of the explanatory weight, maintaining strong, negative, statistically significant relationships with corruption in every model he runs. Electoral democracy shows a positive, if not statistically significant, relationship with corruption in the same models. That is a strange result if one's intuition is that democracy in general is what cleans up government. Sung's own explanation helps explain this dynamic: elections create competition, but competition itself creates new opportunities for vote-buying and influence-peddling, and only produces cleaner governance once paired with an independent judiciary willing to prosecute wrongdoing and a press willing to expose it.

This distinction matters enormously for how this report frames "accountability". It is not democracy as a general condition that converts representation into lower corruption; it is specifically the presence of judicial independence and a free press capable of imposing real cost on corrupt behavior after the fact. Women in political office, often newer to institutions and subject to a heightened, almost disproportionate scrutiny relative to their male counterparts, may rationally judge that the reputational cost of corrupt behavior falls harder on them specifically wherever these factors are active. Thus, not because they are inherently more virtuous, but because judicial and press scrutiny apply asymmetric pressure onto those least protected by existing networks of power.

In systems lacking these two institutions specifically, this logic reverses. Where wrongdoing is unlikely to be prosecuted or exposed, the incentive to avoid it collapses for everyone, regardless of gender. Alatas et al. (2009) found no significant gender gap in corruption tolerance among Indian, Indonesian, or Singaporean samples, while the same experiment produced a clear gap among Australian participants. This contrast is entirely consistent with institutional consideration. Transplant the same individuals into a different incentive environment, and their behavior would naturally shift alongside it. Therefore, corruption tolerance is not something fixed within a person; it is something calibrated, again and again, to the surrounding system of consequences that person happens to be operating in.

I would go further and argue that this is precisely why the popular framing of women as inherently "cleaner" political actors does real analytical damage: it locates the cause of lower corruption inside the individual. More convincingly, it locates it inside the institutions surrounding her, and inside two institutions in particular. Where judicial independence and press freedom are strong, presence changes behavior. Where they are absent or low, presence changes very little at all.

The Structural Exclusion of Women

If accountability explains why individuals behave differently across institutional contexts, it does not explain why women end up underrepresented within corrupt systems in the first place. Bjarnegård's (2013) concept of "homosocial capital" answers this. Her research reframes corruption not as an individual ethical failure, but as an informal system of exchange (favors, information, protection, trust accumulated over years) that depends on ongoing relationships built on perceived reliability. In clientelist political systems, these relationships are overwhelmingly built among men, and not, crucially, because women are morally unwilling to take part in them. Rather, women are not trusted as partners within networks that were constructed, over long stretches of time, by and for men.

This distinction matters enormously when thinking about policy design. If women's lower involvement in corrupt networks reflected some kind of ethical superiority, the policy implication would be straightforward: simply add more women. But if it instead reflects structural exclusion from networks of trust, then adding women to formal institutions without touching the underlying networks does not eliminate those networks. It simply leaves newly elected women standing outside the informal channels through which political business actually gets done. Essentially, the seat and the network are two different things entirely, and holding one does not guarantee access to the other. This can produce a wide range of outcomes, including the marginalization of the very women reformers hoped would change things.

The U4 Anti-Corruption Resource Centre's observation that patronage networks actively hinder women's ability to run for office in the first place extends this logic: exclusion does not begin once women take office, but well before it, in the candidate-selection processes controlled by these same informal networks. This is exactly why Esarey and Schwindt-Bayer's (2019) evidence of bidirectional causality is relevant. Corruption suppresses representation, and not only the reverse. A system organized around trust networks that exclude women will structurally resist their entry, regardless of formal quotas or legal guarantees of eligibility written into law.

Juggling Representation And Accountability

The most consequential insight of this report emerges only once these two mechanisms are combined rather than treated separately. Accountability institutions determine whether individuals already in office face genuine consequences for corrupt behavior. Homosocial networks determine who gets into office in the first place, and through which channels of trust. A reform strategy focused solely on increasing the number of women in office (through quotas, targeted funding, or candidate training programs) addresses only the second mechanism, and even then, only partially. It does nothing to dismantle the informal networks blocking full participation, and nothing to strengthen the accountability structures that would make corrupt behavior costly once someone actually holds office.

This is precisely why representation in weak-institution settings often fails to reproduce the corruption reductions found in cross-country statistical studies. Those studies are disproportionately weighted toward contexts where some meaningful degree of accountability already exists. Applying their headline finding uniformly, as though representation alone were some portable anti-corruption tool that travels across contexts, risks a serious mismatch between diagnosis and setting. In a context marked by weak accountability and strong patronage networks, newly elected women could find themselves facing a difficult choice: remain outside the functioning informal channels and become politically ineffective, or gradually get absorbed into the networks the reform was meant to disrupt in the first place.

4. Policy Recommendations

The evidence assembled above points to a fairly clear conclusion: treating women's political representation as a standalone anti-corruption tool is an error. Its effectiveness depends entirely on the accountability environment it is introduced into and, on two institutions rather than on democratic conditions in general. Electoral competition by itself does little and may even correlate with more corruption where it exists in isolation.

Accordingly, I recommend that international donors and domestic reform coalitions adopt a ‘dual-scope’ approach: gender-representation support (quotas, campaign funding, candidate training) should be paired with two specific institutional investments: strengthening judicial independence and supporting a free, investigative press, rather than the more generic "governance reform" packages. At present, gender programming and these two institutional priorities are funded as entirely separate line items, which is part of the problem.

Stakeholders: This recommendation is aimed primarily at bilateral and multilateral development agencies (UNDP, USAID, EU democracy-support programs, and comparable bodies) and domestic electoral commissions, civil-society coalitions, and women's political networks. Implementation would require genuine coordination between gender-focused governance or anti-corruption program officers within the same institutions, who currently tend to operate in entirely separate funding streams.

Geographic scale and timeframe: This approach is most relevant to countries transitioning toward greater democratic institutionalization, but still characterized by patronage politics. Essentially, a mid-institutionalization where accountability mechanisms exist on paper but remain weak in practice. A realistic implementation timeframe spans a full electoral cycle, typically four to five years, to allow candidate-support programs and accountability infrastructure to mature side by side, rather than the shorter one-to-two-year grant cycles that currently dominate gender-inclusion programming.

Precedent: Rwanda's post-genocide constitutional gender quota, introduced alongside broader governance and anti-corruption reforms during the 2000s, is often cited as a case where representation coincided with governance improvements. Though scholars are right to caution that Rwanda's highly centralized political system limits how far this case can be generalized to more pluralistic settings. A more directly comparable model can be found in Bauhr, Charron, and Wängnerud's (2018) subnational European data, where the corruption-reducing effect of female representation was strongest specifically in council systems that already had stronger transparency norms in place.

Projected outcomes and risks: The dual-scope approach should reduce the risk of newly elected women being marginalized or slowly absorbed into existing patronage networks, and should increase the likelihood that representation translates into measurable governance improvements rather than symbolic change alone. The central risk is donor coordination failure: gender and governance programming are frequently separated with entirely independent metrics of success, and without deliberate integration, this recommendation risks being adopted incorrectly.

Interaction between policy components: The two tracks recommended are complementary rather than merely parallel. A judiciary willing to prosecute corrupt officials and a press willing to expose them reduce the payoff of remaining inside patronage networks for all officeholders. In turn, this lowers the structural barrier facing women seeking entry into political office in the first place. In other words, the second track amplifies the first; it does not simply run alongside it.

Feasibility: This approach requires no major reforms, since judicial independence and press freedom programming already exist as established donor categories in most countries receiving democracy assistance. Thus, the obstacle becomes integration, in this case. The primary obstacle is institutional and administrative. Realigning existing funding streams so that gender-inclusion officers and rule-of-law officers are made to coordinate rather than operate on separate tracks. This creates a feasible reform relative to border institutional changes often proposed and rarely implemented.

Download the full document

References

1. Dollar, D., Fisman, R., & Gatti, R. (2001). "Are Women Really the 'Fairer' Sex? Corruption and Women in Government." Journal of Economic Behavior and Organization, 46, 423–429.

2. Swamy, A., Knack, S., Lee, Y., & Azfar, O. (2001). "Gender and Corruption." Journal of Development Economics, 64, 25–55.

3. Sung, H.-E. (2003). "Fairer Sex or Fairer System? Gender and Corruption Revisited." Social Forces, 82(2), 703–723.

4. Esarey, J., & Schwindt-Bayer, L. A. (2019). "Estimating Causal Relationships Between Women's Representation in Government and Corruption." Comparative Political Studies.

5. Bauhr, M., Charron, N., & Wängnerud, L. (2018). "Close the Political Gender Gap to Reduce Corruption." U4 Brief 2018:3, Chr. Michelsen Institute.

6. U4 Anti-Corruption Resource Centre. "Corruption and Gender Equality." u4.no.

7. Barnes, T. D., & Beaulieu, E. (2014). "Gender Stereotypes and Corruption: How Candidates Affect Perceptions of Election Fraud." Politics & Gender, 10(3), 365–391.

8. Alatas, V., Cameron, L., Chaudhuri, A., Erkal, N., & Gangadharan, L. (2009). "Gender, Culture, and Corruption: Insights from an Experimental Analysis." Southern Economic Journal, 75(3), 663–680.

9. Frank, B., Lambsdorff, J. G., & Boehm, F. (2011). "Gender and Corruption: Lessons from Laboratory Corruption Experiments." European Journal of Development Research, 23, 59–71.

10. Esarey, J., & Chirillo, G. (2013). "'Fairer Sex' or Purity Myth? Corruption, Gender, and Institutional Context." Politics & Gender, 9(4), 361–389.

11. Bjarnegård, E. (2013). Gender, Informal Institutions and Political Recruitment: Explaining Male Dominance in Parliamentary Representation. Palgrave Macmillan.

12. Goetz, A. M. (2007). "Political Cleaners: Women as the New Anti-Corruption Force?" Development and Change, 38(1), 87–105.

Follow our socials

Community

© 2026. All rights reserved.