Does war improve women’s political representation?

Abstract

An expanding body of scholarly work posits that war, particularly intrastate war, disrupts established power arrangements and social institutions, providing avenues for women to enhance their political influence through reforms in political structures and participation in peace-building processes. Webster, Chen, and Beardsley (2019) systematically explore the nexus between war and women’s political empowerment, demonstrating associations between both interstate and intrastate conflicts and advancements in women’s political empowerment. Extending their analysis, this study investigates the influence of intrastate conflicts on women’s representation in legislative bodies. However, nuanced findings emerge from this analysis. There is no supporting evidence that the occurrence of warfare leads to an increase in women’s legislative representation. Ongoing internal conflicts exhibit a negative impact on the progression of women’s political representation. Conversely, the termination of civil wars, particularly without their prompt resumption, correlates with a higher likelihood of adopting gender quotas, thereby enhancing women’s legislative representation. Furthermore, the correlation between the termination of civil wars and women’s legislative representation appears specific to the context of sub-Saharan Africa. No conclusive evidence supports this relationship as a global phenomenon.

How does war affect gender inequality in politics? War consists of several distinct dimensions—“the identification of threat; the mobilization, deployment, and legitimation of violence; and warfare itself” (Ralston and Krebs, 2018: 4). These distinct processes can exert different or even contradictory effects on women’s political empowerment. For example, the expectation of war tends to increase threat perception in a society and its level of militarization. Both processes reinforce traditional gender roles and exacerbate gender power imbalances (Enloe, 1983). Contrarily, periods of warfare can disrupt societies and open a window of opportunity for transformational changes in women’s role in politics (Berry, 2018; Hughes and Tripp, 2015; Tripp, 2015). Thus, different strands of scholarship have produced conflicting expectations about the relationship between war and women’s political empowerment.

However, recent studies demonstrate that war has transformative effects on traditional gender roles and improves women’s political empowerment (Bakken and Buhaug, 2021; Berry, 2018; Hughes and Tripp, 2015; Tripp, 2015). Among others, Webster et al. (2019) (hereafter WCB) provide the most comprehensive and systematic analysis of the relationship between war and women’s political empowerment. Using cross-national data from 1900 to 2015, WCB find that both interstate and intrastate war occurrence are associated with improvements in women’s empowerment in the short- and medium-term. They further show that war improves women’s empowerment by disrupting established social and political orders, increasing the likelihood of regime change, and changing gender roles.

Despite WCB’s important contributions, however, three important puzzles remain unresolved. First, their study cannot determine whether war improves women’s legislative representation because the measure of women’s political empowerment used in their analysis is an aggregate index obtained from the Varieties of Democracy (V-Dem) (Coppedge et al., 2022) (Figure 1). It is the average of Women’s civil liberties, Women’s civil society participation, and Women’s political participation. Women’s political participation is the average of two indicators: Female legislators, the percentage of lower chamber female legislators, and Power distributed by gender, measuring expert-coders’ de facto assessment of the extent to which political power is distributed according to gender. Female legislators is the only observational indicator, while all others rely on expert-coders’ assessment. Importantly, its change does not positively correlate with a change in other expert-coded indicators (see Appendix Section A). Thus, we cannot conclude that war improves women’s legislative representation, although WCB posit that improvements in women’s descriptive representation are a necessary condition for improvements in women’s empowerment. The various components of the women’s empowerment index can differently respond to war.

Figure 1.

Structure of aggregation for the V-Dem Women’s political empowerment.

Next, WCB find little difference between intrastate and interstate wars in improving women’s political empowerment. ¹ This is puzzling, given that intrastate and interstate wars differ in many aspects: intrastate wars tend to be more frequent, deadlier, and longer-lasting. Many scholars thus argue that intrastate war presents unique opportunities to renegotiate and reshape domestic governance and political institutions (Bakken and Buhaug, 2021; Hughes and Tripp, 2015; Tripp, 2015). For the same reason, WCB expect that role shifts caused by war are more likely to occur after the end of civil war than following interstate war. However, their findings do not support this expectation.

Last, WCB find that both ongoing war and war termination increase women’s political empowerment. This finding is inconsistent with other studies arguing that only civil war termination improves women’s political empowerment and representation (Bakken and Buhaug, 2021; Tripp, 2015). Thus, more analyses should be conducted to examine how intrastate warfare and its termination affect women’s legislative representation.

I revisit WCB’s empirical analyses and extend them. My analysis reveals several important findings. First, I find that war occurrence, particularly civil war, is associated with a decrease in women’s legislative representation. WCB’s conclusion does not hold true for women’s legislative representation. Yet, the occurrence of interstate war is associated with its increase. Thus, lumping together civil and interstate war is not useful when scholars explore the consequences of war for women’s political representation.

However, this does not nullify the potential contribution of civil wars to women’s legislative representation entirely. When I examine the effect of civil war termination, I find that civil war termination is associated with an increase in women’s legislative representation and an increased likelihood of gender quota adoption in the short- and medium-term. Nonetheless, this correlation only emerges under the condition that the civil war does not resume in the near future, such as within three or five years. Absent this condition, the effect of civil war termination is statistically insignificant. These results confirm the results found by Tripp (2015) and Hughes and Tripp (2015).

Meanwhile, I find an important qualification: the relationship between civil war termination and women’s legislative representation is confined to the context of sub-Saharan Africa. Contrary to scholars’ expectations, no conclusive evidence supports this relationship as a global phenomenon. Even when I examine other expert-coded indicators related to women’s political empowerment, I find the positive effects of civil war termination on expert-coded indicators, but only in the sample of sub-Saharan Africa. These results, taken together, indicate that women’s parliamentary gains due to the ending of civil war seem to translate into women’s political empowerment in sub-Saharan Africa.

War and women’s legislative representation

To examine the relationship between war and women’s legislative representation, I mostly rely on the data provided by WCB replication files and their model specifications. The dependent variable is the proportion of female legislators obtained from the V-Dem data (version 12). Following WCB’s strategy, I estimate the following linear models:

$\mathrm{\Delta }{y}_{it+s,t-1}=\alpha y_{it-1}+\beta {\text{war}}_{it}+{\mathbf{X}}_{\mathbf{it}-\mathbf{1}}\mathbf{\Gamma }+\mathbf{\Delta }{\mathbf{X}}_{\mathbf{it}-\mathbf{1}}\mathbf{\Omega }+u_i+{ϵ}_{it}$

(1)

where Δy_it+s,t−1 is the change in the proportion of female legislators for country i between years t − 1 and t + s, and s takes one of the following values: 0, 1, 2, 3, 4, 5, 10, and 15. This set-up is suitable for estimating the dynamic response of female legislative representation to war and investigating possible nonlinear effects of war. The independent variable is a binary indicator for the occurrence of either interstate or intrastate wars in a given year, as coded in the Correlates of War for the period of 1900-2006 and the UCDP Armed Conflict Data for the period of 2007-2015. X _it−1 is a vector of country-specific controls: the temporal lag and the change of the Polity IV index and a logged energy consumption variable, the change in women’s legislative representation in neighboring states, and a yearly trend. This model includes country fixed effects u _i and the lagged dependent variable y_it−1.

Panel 1 of Figure 2 plots the coefficient estimates of war and their 95% confidence intervals from eight linear regressions. The result stands in stark contrast to WCB’s result. All coefficients on war are negative, regardless of temporal structures. They are statistically significant in the short run but are statistically insignificant in the medium or long run. Following WCB, Panels 2-4 differentiate between new, ongoing and recent war to further explore temporal dynamics. ² WCB find that both ongoing and recent war are, respectively, associated with medium-term and short-term improvements in women’s empowerment. However, even when I differentiate between new, ongoing, and recent war, little evidence suggests that war is associated with improvement in female legislative representation. Only coefficient estimates on recent war are positive in the models of long-run change but are statistically insignificant.

Figure 2.

Marginal effects of war on female legislative representation. The graph plots coefficient estimates with 95% confidence intervals for regression models. A dependent variable is a change of female legislative representation from t − 1 to t + s, and s is displayed at the x-axis of each graph.

A similar result emerges when I use war duration or battle deaths instead of the war dummy. If wars serve as a catalyst for improving female political empowerment by disrupting preexisting social hierarchies, their effect should be greater when wars are more intense, costly, and prolonged (Bakken and Buhaug, 2021; Hughes and Tripp, 2015). However, Panels 5-6 show that both war duration and battle deaths are negatively associated with the percentage of female legislators. Particularly, coefficient estimates on battle deaths are statistically significant in most models.

To answer why I find different results from WCB, I estimate the effect of war on changes in women’s empowerment and its components. I use the same data used in WCB (V-Dem version 6.2) for the purpose of comparison. Figure 3 shows that women’s political empowerment and expert-coded components, except Women’s civil liberties, exhibit a similar pattern: coefficient estimates on war are mostly positive and their magnitude increases over time and decline later. ³ Particularly, a stark difference between the two sub-components of Women’s political participation stands out. Unlike those of war on female legislative representation, the effects of war on Power distributed by gender are positive and statistically significant. Their dynamic pattern is very similar to the pattern of the aggregate empowerment index. This suggests that WCB’s results are mainly driven by Power distributed by gender.

Figure 3.

Effects of war on female political empowerment and its components.

I further examine the effects of irregular leadership changes or regime changes on the same dependent variables since WCB conclude that regime change, measured by irregular leadership change, is the most feasible intermediate mechanism. Figure A10 shows that irregular leadership changes are associated with improvements in Power distributed by gender and Women’s civil participation. A similar pattern emerges when I directly examine the effect of democratic transitions (Figure A11). Taken together, these findings suggest that in WCB’s results, war enhances women’s political empowerment through Power distributed by gender and Women’s civil participation by catalyzing democratic transitions. However, war does not have positive effects on women’s legislative representation since democratization does not increase women’s legislative representation (Fallon et al., 2012).

Distinguishing between intrastate and interstate war

I distinguish between intrastate and interstate wars. Panel 1 of Figure 4 shows that intrastate and interstate wars differently affect female legislative representation. The coefficients of intrastate wars are negative and statistically significant at the 5% or 10% level in most models. Substantively, intrastate wars are associated with a decrease in the percentage of female legislators by 0.4 to 1.5 percentage points. Conversely, the coefficients on interstate wars are mostly positive but statistically insignificant. Panels 2-4 differentiate between new, ongoing, and recent intrastate wars. The effects of ongoing intrastate wars are negative, while those of ongoing interstate wars are mostly positive. I fail to find that new or recent wars, regardless of types, improve female legislative representation. Last, Panels 5-6 report that both intrastate war duration and battle deaths are correlated with a negative change in female legislative representation. They suggest that more intense and protracted civil wars more negatively influence female legislative representation. I find little evidence that intrastate wars contribute to the growth of female legislative representation. ⁴ The results indicate the importance of distinguishing between intrastate and interstate wars.

Figure 4.

Distinguishing between intrastate and interstate wars.

Intrastate war termination and female legislative representation

Previous results seem to challenge the existing studies demonstrating a positive relationship between civil war and women’s legislative representation. ⁵ However, it is important to note that unlike WCB, other scholars underscore the role of civil war termination in fostering the growth of female political representation (e.g., Bakken and Buhaug, 2021; Hughes and Tripp, 2015; Tripp, 2015). They argue that ongoing civil wars negatively affect female political representation. Particularly, Hughes and Tripp (2015) demonstrate that civil war termination, not ongoing civil war, improve women’s legislative representation in sub-Saharan Africa if war does not resume.

Thus, I examine the relationship between intrastate war termination and female legislative representation. I construct a binary indicator for civil war ending where the previous year experienced intrastate war but the next 3 years did not. I include it in the previous models used in Figures 2 and 4 and also control for ongoing civil wars. Figure 5 shows the importance of differentiating between ongoing war and war termination. Civil war ending is positively associated with an improvement of women’s legislative representation, while ongoing civil war is negatively associated with it. ⁶

Figure 5.

Effects of ongoing intrastate wars and intrastate war termination on women’s legislative representation. Intrastate war ending refers to the situation where the previous year experienced war but the next 3 years did not.

Next, I limit the sample temporally or geographically, as in Hughes and Tripp (2015). ⁷ When I limit the sample to years since 1985 or sub-Saharan Africa, the effects of civil war termination increase and are statistically significant in the medium run. However, the same cannot be said of regions outside sub-Saharan Africa: the effects of civil war termination are close to zero (see also Tables A1-A4). Similar results emerge when I examine the relationship between civil war termination and the adoption of gender quotas. Figure A18 indicates that civil war termination is positively associated with an increased likelihood of gender quota adoption, but only in sub-Saharan Africa. These results challenge Hughes and Tripp’s expectation that their finding will hold in other regions.

Last, I estimate the effect of ongoing civil war and civil war termination on changes in women’s empowerment, its three components, and the two sub-components of women’s political participation. ⁸ Based on the previous results, I also split the sample between sub-Saharan Africa and other regions in Figure 6. In sub-Saharan Africa, civil war termination is positively associated not only with the growth of women’s legislative representation but also with an increase in expert-coded indicators, such as Women’s civil liberties and Power distribution by gender. This is important since women’s descriptive representation in the legislature does not necessarily translate into empowerment. Many autocrats strategically increase the share of female legislators to increase domestic and international support for the regime (Bjarnegård and Zetterberg, 2022). Female legislators may not exercise real policy-making power. However, the finding that civil war termination produces improvements in both observation and expert-coded indicators can assuage the concern about women’s legislative representation. Women’s parliamentary gains in the wake of civil war seem to translate into women’s political empowerment in sub-Saharan Africa.

Figure 6.

Effects of war on expert-coded components of female political empowerment. Each dependent variable is measured as a change of each indicator from t − 1 to t + s, and s is displayed at the x-axis of each graph.

Conclusion

By revisiting and extending the WCB’s research, this paper confirms that civil war termination can inadvertently contribute to women’s political representation. Particularly when war does not resume soon, the termination of civil war is associated with an increased likelihood of gender quota adoption and an improvements in women’s legislative representation. On the other hand, internal warfare itself negatively affects the growth of women’s political representation. However, this paper provides an important qualification: the relationship between civil war termination and women’s political representation and empowerment is limited to the context of sub-Saharan Africa. No positive effects of civil war termination are found beyond sub-Saharan Africa.

These findings provide an important implication for studying the consequences of war. It is important to treat war as a process rather than an event and disaggregate it into distinct constituent processes (Ralston and Krebs, 2018). Distinct processes can exert different effects on women’s political representation and empowerment. Similarly, the finding that intrastate and interstate war differently influences women’s political representation indicates that scholars should be careful about lumping intrastate and interstate war together.

Last, this paper has important implications for using the women’s political empowerment index or its subjective component. Different component indicators differently respond to war, war termination, and democratic transitions. Using the aggregate index can hide different patterns behind each indicator. Scholars should also be cautious about making broad claims based on one specific indicator. At the same time, it is important to check what an expert-coded indicator measures. For example, Power distributed by gender neither correlates much with female legislative representation nor increases after the adoption of gender quotas (see Figures A5 and A6).

This may be because it reflects expert-coders’ perceptions about power distribution by gender beyond the legislature. Thus, an increase in female legislative representation does not necessarily translate into an increase in the subjective measure. However, an outstanding question remains regarding what the indicator captures and what determines it (see Appendix Section F). Scholars should try to validate this indicator and examine the possibility that experts’ bias comes into play before they use it for empirical analysis.

Supplemental Material

Supplemental Material - Does war improve women’s political representation?

Supplemental Material for Does war improve women’s political representation? by Nam Kyu Kim in Research & Politics