1 Introduction

Since the escalation of the COVID-19 pandemic in March 2020, the role of female leadership in national governments has captured significant interest from both the media [1, 2] and the academic community [3,4,5,6,7,8,9,10,11,12,13]. Narratives and empirical studies suggest that female heads of state, such as presidents and prime ministers, have demonstrated effective leadership, prompt responses, and the implementation of inclusive policies, leading to reduced mortality rates during the pandemic [3, 7, 9, 11].

Despite these findings, some scholars raise questions about the validity of these correlations. Initially, the historical and global prevalence of female leaders in top governmental positions has been limited [14]. As of March 2020, only eleven women occupied these leading roles [6, 7].

Additionally, the presence of significant female leadership is disproportionately concentrated in certain regions, primarily in the more affluent Global North. These regions typically exhibit higher GDPs, lower levels of governmental corruption, and greater public health expenditures compared to others [6, 7, 15]. The cross-sectional analysis, often employed in this field, does not adequately control for various confounding factors. Consequently, one of the studies contends that the apparent positive correlation between female leadership and effective pandemic management is “spurious” [6].

Lastly, while the focus has predominantly been on the gender of government leaders and their pandemic responses, there is a notable scarcity of research exploring the role of legislative bodies in this context [4, 12].

In an effort to mitigate these shortcomings, this study explores the impact of female representation in national legislature on epidemic response efficacy. We employ two-way fixed effect (TWFE) and instrumental variable (IV) regressions to evaluate epidemic responses over the last century. While recent studies predominantly focus on female leadership during the COVID-19 pandemic, there is a comparative lack of research on previous epidemics. TWFE models provide a more robust framework for this investigation. However, these models are not without their limitations, particularly concerning unobserved heterogeneity. To address this, we utilize TWFE and IV models to establish a causal relationship. Our analysis demonstrates that the role of female legislators is more pivotal than that of female chief executives in reducing epidemic-related mortality. Furthermore, our extended analysis suggests that the proportion of female legislators, in conjunction with the deliberative process, plays a significant role in diminishing death rates during epidemics.

This study also contributes to the literature on politics and gender. The literature examines the relationship between politicians’ gender and the policies they implement [16, 17] and people’s health [18,19,20]. However, epidemic deaths are not sufficiently covered in the literature, especially those analyzed with historical datasets. This study offers new evidence by applying a historical dataset and a causal inference approach to this issue.

2 Gender, policymaking, and epidemic

The academic discourse extensively explores the nexus between the gender of politicians and policy outcomes. This scholarship probes the impact of gender on various domains such as policy topics [16, 17], government spending [21,22,23,24], people’s health [18,19,20], and educational initiatives [25].

Additionally, quotas can affect policy responses. First, quotas increase the presence of female politicians, which affects policy topics, government spending, people’s health, and educational initiatives, as discussed. Second, quotas change the political culture, which affects policy choices. Quotas expose the legislators to women’s organizations and policy areas familiar to female politicians [26,27,28]. Thirdly, gender quotas are justified independently of policy-output reasons [29, 30].

Furthermore, although several scholars indicate the simple correlations between female representation and positive policy output, as discussed before, other scholars insist that there should be a threshold where female presence starts to play a significant role–or critical mass– in the legislature. The proponents of the critical mass theory insist that only a few percent of female members of an organization are subject to greater visibility within the group, leading dominants to stress intra-group differences. As a result, minorities are compelled to conform to dominant models while also suffering stereotypes [31, 32]. The theory is also examined in the political contexts recently [24, 33, 34].

Building on the literature on gender quotas and the critical mass theory, it is suggested that a certain proportion of female representation is necessary for women to exert significant influence in politics [35,36,37,38]. At the threshold, female legislators start to work more effectively to influence policy outcomes, challenge the status quo, and ensure that women’s perspectives are integrated into the legislative process.

Extending the literature on female representation and politics, recent empirical studies posit that female leadership in both national and local governmental spheres has been notably effective during the COVID-19 pandemic [3,4,5,6,7,8,9, 11,12,13, 39]. Additionally, media narratives have emphasized the significant contributions of prominent female leaders during this crisis, including Jacinda Ardern of New Zealand, Mette Frederiksen of Denmark, and Sanna Marin of Finland [1, 2].

Yet, the literature scarcely addresses the causal relationship between the gender of politicians and their effectiveness during epidemics, except in a limited number of cases [15, 39]. This paucity of research stands in stark contrast to the extensive body of work on the influence of politicians’ gender in policymaking, which frequently adopts causal inference methodologies such as natural experiments [21], regression discontinuity [25, 40, 41], and the application of instrumental variables [17, 42, 43].

Furthermore, despite existing studies exploring the link between political institutions and HIV/AIDS through panel data analysis [44, 45], there remains a notable gap in research addressing the relationship with epidemics more broadly, barring a few exceptions [46]. This gap is particularly noteworthy in light of the historical global focus on epidemics prior to the advent of COVID-19 [47].

This research employs century-long epidemic mortality data alongside the instrumental variable methodology to probe a critical inquiry. The study first assesses how the gender of politicians influences epidemic mortality rates. Predominantly, women exhibit distinct policy predilections compared to men. Women, generally more risk-averse, particularly in health-related contexts [48, 49], have demonstrated proactive approaches in pandemic management, as evidenced in COVID-19 research [10, 15]. In addition, female politicians are typically inclined toward prioritizing sectors traditionally associated with women’s interests– healthcare, social services, family and child welfare, and environmental concerns, as discussed before. These tendencies indicate a female political focus on preventive strategies in epidemic scenarios.

Furthermore, women’s leadership frequently prioritizes cooperative and collaborative methods over traditional hierarchical control. This leadership style fosters an inclusive and participatory decision-making process. Such an approach has been linked to better outcomes during health crises [50, 51]. Specifically, collaborative leadership style often adopted by women is associated with reduced fatalities during epidemics [5, 9]. This link suggests that the emphasis on teamwork, shared responsibility, and transparent communication inherent in female leadership may contribute to more effective and timely responses to public health emergencies, thereby mitigating the impact of infectious disease outbreaks.

When the share of female politicians increases, their distinctive perspectives and priorities are more likely to be reflected in the policymaking process. Women in politics tend to emphasize cooperative and collaborative methods over hierarchical control, show heightened interest in healthcare and other traditionally female policy areas, and exhibit greater risk aversion. Quotas can send policy cues to both male and female policymakers about women’s interests and priorities during the policy adoption process [26]. Furthermore, the increase in female representation raises the probability of female politicians attaining high-prestige cabinet posts [52, 53], which enables them to exert greater leadership influence. In summary, an increase in female politicians allows their interests and risk-averse nature to be more effectively integrated into the policymaking process and promotes a leadership style characterized by cooperation and collaboration [16, 17, 54]. These factors collectively contribute to a reduction in epidemic-related fatalities.

Lastly, minority representation in political discourse, the representation of minority groups (such as ethnic or sexual minorities) in the decision-making processes of governance, ensures the inclusion of diverse viewpoints. Epidemic prevention strategies often intersect with various rights, particularly those of minority groups [55]. Majority group politicians may favor policies that minimally impact their own rights. Conversely, leaders representing minorities typically advocate for more inclusive approaches [3]. Female legislators, for example, demonstrate reluctance towards endorsing disruptive measures like school closures, considering the broader familial implications [12]. Additionally, female leaders often emphasize the impacts on individuals and families, eschewing bellicose rhetoric [56]. Therefore, we can expect that the increase in female politicians urges the government to implement more inclusive policies, which leads to decrease of epidemic deaths.

As a result,


Hypothesis 1: When the rate of female legislators is high, epidemic deaths decrease.


Hypothesis 2: When the head of the government is female, epidemic deaths decrease

3 Data and methods

This research aims to delineate the causal dynamics between female political leadership and epidemic mortality rates. We source our dependent variable from the public database, EMDAT [57], which exhaustively catalogs various disasters and their fatalities by country from 1900 to 2019 [46]Footnote 1. This database has been instrumental in a multitude of prior studies [46, 58, 59]. The database includes the variable reporting the total number of deaths related to each disaster. This column includes confirmed fatalities directly imputed to the disaster plus missing people whose whereabouts since the disaster are unknown and so they are presumed dead based on official figures [57]. In addition, EM-DAT classifies disasters into several categories. We extract those classified as “Epidemic” for the following analysis. We construct country-year data on infections and fatalities by transforming the EM-DAT database, which records incidents of infectious disease disasters by country alongside fatalities. The dependent variable, Epidemic Deaths per 1 M, quantifies annual epidemic fatalities per million individuals in each countryFootnote 2.

However, the data’s reliability prior to 1988 is compromised due to data collection inconsistencies [57]. Aligning with the literature [46], we ensure robustness in our analysis by excluding pre-1988 data (Appendix Table 5).

The primary independent variable is the gender of the head of government, derived from V-dem [60] (HOG female). The secondary independent variable, the proportion of female legislators in the lower legislative chamber, is also obtained from V-dem [60] (Share of Female Legislators). Control variables incorporated into our analysis include Gross Domestic Product (GDP), population size, and the Liberal Democracy Index, as enumerated by V-dem [60].

To examine the relationship between politicians’ gender and epidemic deaths, we employ two-way fixed effect (TWFE) and instrumental variable (IV) regressions. While recent studies predominantly focus on female leadership during the COVID-19 pandemic, there is a comparative lack of research on previous epidemics. TWFE models provide a more robust framework for this investigation. However, these models are not without their limitations, particularly concerning unobserved heterogeneity. To address this, we also utilize IV models to establish a causal relationship.

The instrumental variable, YearsWS, quantifies the duration since the inception of women’s suffrage, drawing upon V-dem [60]Footnote 3. This metric has been previously utilized to assess the impact of female legislative representation on policy dynamics [17, 42, 43].

An IV approach can enable causal identification provided that our instrument (YearsWS) is sufficiently correlated with the endogenous variable (Share of Female Legislators) and only influences the dependent variable (Epidemic Deaths per 1 M) through its effect on that endogenous variable. Our IV meets both criteria. First, the first-stage analysis with OLS revealed a significant correlation between YearsWS and Share of Female Legislators (Appendix Table 2). The F-statistic (7576) from the first stage of our two-stage least squares regression exceeds the commonly accepted threshold of 10, indicating that weak identification is not a concern.

Second, there is no obvious alternative pathway through which the years since the suffrage of women affect epidemic deaths that do not involve female representation in the legislature. A potential concern pertained to whether women’s suffrage might inadvertently influence life expectancy, thereby affecting epidemic mortality rates, potentially breaching the exclusion restriction. We investigated this by analyzing the association between YearsWS and life expectancy across countries and years through TWFE models. The findings indicated no substantial correlation, validating the appropriateness of YearsWS as an instrumental variable (Appendix Table 3).

In sum, YearsWS is plausibly exogenous, making it a good instrument for the female representation.

As a result, we compiled a dataset of 183 countries for 123 years. Summary statistics of the variables are shown in Appendix Table 1. Although the comprehensive dataset provides a robust basis for analysis, the power of our analysis was found to be low, at 0.05. This low power is due to the constraints of observational data, limiting the sample size and thereby the ability to detect smaller effect sizes.

4 Results

Initially, we deployed TWFE models to scrutinize the correlation between the gender composition of political leaders and epidemic mortality rates. Model (1), as delineated in Table 1, indicate that Hypothesis 1 lacks empirical support. Model (2), even after the inclusion of various control variables, echoes this finding, revealing no significant impact of the gender of the head of government on epidemic fatalities. Conversely, Models (3) and (4) affirm Hypothesis 2, evidencing a negative correlation between the proportion of female legislators and epidemic death rates.

Table 1 Fixed Effect Models
Full size table

To bolster the causality and reliability of our findings, we addressed potential confounding factors of unobserved heterogeneity. We adopted an instrumental variable approach, utilizing a two-stage least squares (2SLS) model. The instrumental variable analysis results, presented in Table 2, consistently demonstrate a statistically significant association between Share of Female Legislators and Epidemic Deaths per 1 M. These findings suggest that an increase in female legislative representation correlates with a decrease in epidemic-related mortality rates.

Table 2 IV results
Full size table

In addition to the IV analyses, we executed a series of robustness checks. The first pertained to outliers. Because epidemic is sometimes concentrated in certain regions, outliers can significantly affect the results. Therefore, we removed the top ten outliers from the panel data and implemented the analysis. Appendix Table 4 shows that the results are consistent with our primary results.

Secondly, addressing data quality concerns before 1988—highlighted in the Data section and following the precedent [46]—we conducted a robustness analysis by omitting data preceding 1988. The outcomes of this revised analysis, as displayed in Appendix Table 5, are mostly congruent with our primary results, thereby substantiating the stability and reliability of our study’s conclusions.

5 Discussion

5.1 Why share of female legislators?

The preceding analysis delineates that the proportion of female legislators is significantly inversely correlated with epidemic mortality rates, while the gender of the head of government does not demonstrate a significant association. This outcome contrasts with recent studies on the COVID-19 pandemic, which suggested superior performance by female leaders compared to their male counterparts, both in quantitative [15, 39] and qualitative terms [3, 7, 8]. Yet, our findings corroborate research underscoring the critical role of female legislative representation [4, 12]. Employing TWFE and IV methodologies with a century’s worth of epidemic mortality data, our study substantiates the significance of female presence in legislatures.

The divergence in these findings may be attributable to the distinct functions of legislative and executive roles. Legislators are generally expected to represent diverse interests [61], with individual legislators focusing on varied policy areas [62, 63]. This enables female legislators to represent women’s interests by advocating for women-centric policies [54, 64].

In addition, to effectively decrease epidemic deaths, governments must be adequately prepared for a pandemic before it occurs by investing in robust public health infrastructure. This includes investments in health insurance, medical experts, and sanitation infrastructure, which are critical components of a comprehensive epidemic response strategy [65, 66]. Such investments are typically dependent on legislative action in most countries, as they require substantial budget allocations. Therefore, the role of the legislature becomes crucial in ensuring that the necessary resources are allocated to public health preparedness. This proactive approach to public health can mitigate the impact of a crisis more effectively than reactive leadership during an emergency.

Furthermore, female legislators are likely to prioritize public safety during a pandemic. As a result, they can exert pressure on the executive branch to utilize the established infrastructure, such as hospitals, medical professionals, and sanitation systems, to prioritize public safety. The presence of female legislators plays a crucial role, especially in times of crisis, in promoting decision-making that considers the overall welfare of society [18,19,20]. Their active advocacy for public health policies ensures that the legislature fulfills its important role in overseeing the executive branch [67], ensuring that the necessary measures are taken to safeguard the population during emergencies.

In contrast, executive leaders, such as prime ministers and presidents, are more likely to address broader national issues encompassing the economy and foreign affairs. Consequently, it remains ambiguous whether a female executive should, or indeed does, explicitly represent women’s concerns [14]. This scenario suggests that the executive branch maintains more traditionally masculine traits compared to other political branches [68], despite female legislators also confronting gendered challenges within legislatures [69]. Hence, gender differences may not directly translate into distinct policymaking styles at the executive level. This is exemplified by New Zealand’s Prime Minister Jacinda Ardern, lauded for her preventive approach in 2020 [1, 2], yet facing calls for resignation amid economic difficulties in 2023 [70].

The pandemic has not only tested the government’s preparedness prior to its outbreak, but it has also highlighted the importance of diverse legislative representation. Increasing the number of women in legislatures is a more significant and challenging task than electing a female prime minister or president. This complexity arises because while female leaders in executive roles are often selected by male-dominated political structures to maintain existing power dynamics [14, 71], true increases in female legislative representation, exceeding the critical mass threshold [33, 34], can fundamentally alter these power structures, leading to more substantial changes in governance. Thus, when assessing a government’s performance during a pandemic, the proportion of women in the legislature is a more crucial factor than the gender of the head of government.

6 Mechanism

Subsequently, our investigation delved into the mechanisms by which female legislators mitigate epidemic mortality rates. Existing scholarship posits that female leadership was instrumental in the effective management of COVID-19, primarily through the adoption of socially inclusive policies and community-centric approaches [6, 8]. To probe this dynamic, we incorporated variables indicative of the deliberation process and examined their interaction with the share of female legislators. For this purpose, we source Respect for Counterarguments, Reasoned Justification, Range of Consultation, and Engagement with Society from V-dem [60]. Experts assessed the extent to which political elites acknowledge and respect counterarguments, provide public and reasoned justifications for their positions, consult with other elites, and how wide and independent public deliberation is for significant policy changes. According to the literature, female legislators reduce epidemic deaths, particularly when the country exhibits a robust deliberation process where political elites consider and respect opposing views, provide clear rationales, consult other leaders, and where public discourse on significant policy changes is broad and independent. The findings, presented in Tables 3 and 4 and illustrated in Fig. 1, suggest that the share of female legislators interacts with the deliberation process to decrease epidemic deaths.

Table 3 Mechanism 1
Full size table
Table 4 Mechanism 2
Full size table
Fig. 1
figure 1

Share of female legislators interacts with deliberation process to decrease epidemic deaths

Full size image

This empirical evidence reinforces the premise that policies grounded in diversity and inclusion are pivotal in epidemic containment efforts. Marginalized communities typically bear a disproportionate brunt of epidemic crises [72, 73], and certain crisis-driven policy decisions exacerbate their plight [74,75,76]. Higher participation of various minorities in the policymaking process allows the government to implement anti-pandemic policies favoring minorities. Therefore, the incorporation of diverse perspectives in policy formulation and the active engagement in deliberations with minority groups emerge as key components in the effective management and mitigation of epidemic challenges.

7 Critical mass theory

With TWFE and IV models, we show that the increase in the share of female legislators is associated with the decrease in epidemic deaths. However, the proponents of critical mass theory could question the validity because they believe that the critical mass threshold would significantly affect policy outputs.

To examine this theory’s validity, we plot the share of female legislators and the number of epidemic deaths in Fig. 2. This scatter plot illustrates the relationship between the share of female legislators and the number of epidemic deaths per 1 million people (log-transformed). Each point represents a country. The black line represents the smoothed trend line, providing a visual summary of the overall relationship. Notably, the relationship between the share of female legislators and epidemic deaths per million people appears to change around the 15% mark. Initially, as the share of female legislators increases, the number of epidemic deaths also increases. However, once the share of female legislators exceeds approximately 15%, this trend reverses, and further increases in the share of female legislators are associated with a decrease in epidemic deaths per million people.

Fig. 2
figure 2

This scatter plot illustrates the relationship between the share of female legislators and the number of epidemic deaths per 1 million people (log-transformed). Each point represents a country, with the proportion of female legislators on the x-axis and the logarithmic scale of epidemic deaths per million on the y-axis. The gray points indicate individual data points, while the black line represents the smoothed trend line, providing a visual summary of the overall relationship.

Share of female legislators and epidemic deaths per 1 M.

Full size image

The result suggests a potential non-linear relationship between these two variables, which aligns with the critical mass theory. According to this theory, a minimum threshold—in this case, around 15%–is necessary for the presence of female legislators to start having a significant impact. Then, female members of an organization are seen as potential allies and start to increase their influence [31,32,33].

8 Conclusion

In this research, we ascertain that an increased proportion of female legislators correlates with reduced epidemic mortality rates, while the gender of the executive leader does not demonstrate a significant association. This conclusion is drawn from analyses using TWFE and IV models. Further, our investigation into the interaction with various deliberation related variables reveals that their impact diminishes with increasing values, underscoring the efficacy of diversity and inclusion-centric policies in addressing epidemics. Then, we show that such an effect follows the critical mass theory. When the share of female legislators surpasses a threshold, they start to impact the policy output [37, 38].

These findings are pertinent to ongoing discussions concerning optimal policymaking strategies in the context of epidemic response. More broadly, they suggest that heightened female participation in political processes during crises fosters a more balanced decision-making framework. Such an approach more effectively weighs the diverse potential impacts of policy interventions, ultimately serving the collective interests of society.

There are several limitations to our study. First, the availability and quality of data regarding past epidemics pose challenges, particularly in developing countries and historical datasets where information may be incomplete.

Secondly, while instrumental variable analysis mitigates endogeneity concerns, its efficacy hinges on the validity and strength of the chosen instruments. Despite rigorous checks for IV robustness, there remains a possibility that unobserved variables could violate the exclusion restriction assumption, potentially biasing the estimated effects.

Thirdly, we only examine the direct relationship between the share of female politicians and policy outputs. Our discussion does not conclude other paths through which the share affects the output [35, 36] or the importance of female representation or a quota itself [29, 30].