Gender Disparities in Post-Conflict Societies: A Cross-National Analysis

ABSTRACT

This article investigates the impacts of conflict, resolution, and post-conflict democracy on gender bias. Exploring this question poses two methodological concerns. The first regards selection bias in which countries experience conflict and its resolution. The study addresses this issue using a generalization of the Heckman procedure. The second is that post-conflict democracy is likely endogenous to the level of pre-conflict democracy. This issue is addressed using two-stage least squares. Results show that conflict unambiguously worsens gender outcomes with respect to secondary school enrollment, labor force participation, fertility, and parliamentary representation. However, it does not affect the gap in life expectancy. Conflict resolution improves gender outcomes significantly, but not always by a magnitude that restores pre-conflict levels of equality. Greater post-conflict democratization improves parliamentary representation of women and the gender gaps in life expectancy and secondary school enrollment. However, it worsens the gap in labor force participation.

HIGHLIGHTS

• The study corrects selection bias in conflict and its resolution with a three-step procedure.

• It instruments for post-conflict democratization using legal origin and geography.

• Conflict worsens gender inequities in education, the labor force, and representation.

• Conflict resolution mitigates most conflict-induced inequities, but not fully.

• Democratization further improves equity in representation and schooling.

KEYWORDS:

• Gender
• institutions
• equality
• democracy

JEL Codes:

• D74
• J16
• P48

ACKNOWLEDGMENTS

We would like to thank Susan Jellisen for her helpful comments during the early stages of the research, as well as participants in sessions at the 2013 meetings of the International Studies Association.

SUPPLEMENTAL DATA

Supplemental data for this article can be accessed at https://doi.org/10.1080/13545701.2021.1901128.

Notes

1 The full text is available at http://www.un.org/en/ga/search/view_doc.asp?symbol=S/RES/1325(2000).

2 Ghobarah, Huth, and Russett (2003) identify the following mechanisms. First, civil war exposes the civilian population to conditions that increase the risk of adverse health outcomes. Second, it reduces the public expenditure on health, both by eroding the revenue base of the post-conflict regime and by confronting the regime with a competing set of investment priorities, including economic infrastructure, the administrative capacity of the state, the judicial system, and the security infrastructure. Finally, it reduces the efficiency of public expenditure by destroying surveillance and control programs for infectious diseases; the transport, communications, and public-distribution infrastructure necessary to implement basic preventive policies and deliver necessary first responses; and the pool of qualified medical professionals needed to deliver treatment.

3 See Mayra Buvinić et al. (2013) for a survey of the evidence.

4 This, of course, is in addition to the bias that arises from missing or underestimated data due to the collapse of information collection systems at the time of conflict. To the best of our knowledge, the study by Li and Wen (2005) remains the only paper in the literature to explicitly model this or for that matter, any form of selection bias.

5 While even a cursory review of the literature is beyond the scope of this paper, three mechanisms have received scholarly attention: First, SVAW is used as part of the spoils of war awarded to combatants to retain their loyalty (Card 1996). Second, it is used as a tool by organizations to forge a cohesive army out of a collection of individuals, the majority of whom have no social ties to the others and many of whom were coerced into joining (Cohen 2013). In such situations, participation in gang rape is interpreted as a signal that the individual identifies with the group and is willing to incur significant costs in terms of the emotional toll of forced sexual contact and the risk of sexually transmitted diseases to remain part of it. Finally, the rape of “enemy” women serves a dual purpose. Given the subordinate position accorded to femininity in the ideology of war, it acts as an effective tactic for undermining the masculinity of enemy men by proving them to be incapable of protecting their women (Enloe 2000). At the same time, given the symbolic association of women as the bearers of a particular group identity through their reproductive roles, the rape of enemy women destroys the very identity of the rival community. However, it needs to be emphasized that the traditional dichotomy of men as perpetrators and women as victims is questionable, since women can and do participate in atrocities against other women. A survey conducted in the Democratic Republic of Congo in 2010 reports that 41 percent of female victims of sexual violence and 10 percent of male victims were brutalized by female perpetrators (Johnson et al. 2010). See Dara K. Cohen (2013) for a discussion.

6 This includes the categories internal armed conflict and internationalized internal armed conflict in Version 4.0 of the Armed Conflict Dataset of the Uppsala Conflict Data Program (UCDP). See Lotta Themnér and Peter Wallensteen (2013) for more information on the data.

7 We of course realize that there can be a lot of variation in the modes and motivations for conflict. While we control for this partially by using an ethnic polarization in our specification, a fuller treatment of this problem would consider the multiple dimensions of conflict together as separate treatments. Unfortunately, this sort of exercise is not feasible under the identification conditions since it would increase the dimensions of the selection problem geometrically.

8 This is consistent with the current praxis (Bang and Mitra 2017).

9 About fifty-nine incidents of conflict occur in thirty-two different countries within our sample period (depending on the definition of a “new” conflict); another twenty-seven experienced conflict that had resolved prior to our sample period.

10 The excluded components are PARREG and PARCOMP, which respectively measure the regulation and degree of competition in political participation. See James R. Vreeland (2008) on the methodology. The data are available at: http://faculty.georgetown.edu/jrv24/goldindex.html

11 We also replicate our results using the ethnic polarization index, which captures the conflict potential of ethnic divisions (Montalvo and Reynal-Querol 2005), and the geographic variation in the PDSI within a country to capture regional disparities in income shocks. Adding these instruments increases the variance of our selection equations and renders the group of excluded instruments weak (as determined by the joint c²-test). We have included these results as a supplemental appendix (see the Supplemental Materials tab in the online article) and will also make them available on request.

12 The choice of the instrument is motivated by the idea that energy consumption is typically taken as a measure of infrastructure, which is a key determinant of growth (Sahu and Dash 2012). Additionally, there is compelling evidence that energy consumption, in and of itself, may provide an impetus to growth.

13 The joint χ²-statistic for the excluded instruments is 6.41 with a p-value of 0.0406.

14 The joint χ²-statistic for the excluded instruments is 43.04 with a p-value less than 0.0000.

15 The joint F-statistic for the excluded instruments is 26.49 with a p-value less than 0.0000.

16 It needs to be emphasized that the selection equations are not meant to infer causal patterns in conflict or democracy, and it may well be the case that countries with British legal origin face lower rates of post-conflict democratization precisely because they were more likely to have more democratic institutions, higher economic development, and the like in the first place.

Additional information

Notes on contributors

Aniruddha Mitra

Dr. Aniruddha Mitra is Assistant Professor of Economics at Bard College and Research Fellow with the Global Labor Organization in Bonn, Germany. He teaches courses on microeconomics, development economics, and migration. Dr. Mitra’s research focuses on the causes and consequences of conflict for economic development. This research has investigated the role of ethnicity and discrimination in fomenting conflict, as well as the dual roles of globalization and political instability in perpetuating racial, ethnic, and gender bias.

James T. Bang

Dr. James Bang is Associate Professor of Economics at St. Ambrose University and Research Fellow with the Global Labor Organization in Bonn, Germany. He teaches courses on international economics, institutional economics, and econometrics at both the undergraduate and graduate levels. Dr. Bang’s research focuses on institutional determinants of economic development. In this vein, Dr. Bang has published and has forthcoming articles on the determinants of civil war, the impact of civil war on economic growth, and the impact of civil war on cultural attitudes toward women. He has also investigated the role of institutions in determining migration; remittances; and gender norms.