Global Termination and Recurrence Macro Trends: A Follow-Up to Licklider and Dixon

ABSTRACT

In this Special 25th Anniversary Issue, we (re)examine several macro historical trends concerning civil war termination and recurrence that were raised in the journal’s first issue by Roy Licklider and then followed-up on 10 years later by Jeffrey Dixon. These trends have to do with: (1) the global prevalence of recurrence; (2) the global prevalence of termination types; (3) patterns of recurrence by termination type, and (4) post-conflict patterns of civilian victimisation by termination type. With the advantage of more comprehensive and more systematised data, the analysis identifies several structural breaks in termination and recurrence trends, as well as differences in how outcomes regulate the inflow and outflow of fighting groups in countries. Over decades, conflict recurrence has been becoming more and more concentrated in a smaller subset of countries, and the majority of conflict recurrences are caused by new groups and not by previously terminated groups. Around 59 per cent of recurrences following peace accords are caused by new groups with no prior appearance in the conflict record, while 71 per cent of recurrences following government victory are caused by new groups. Since 2010, termination rates have declined by 25 per cent and recurrence rates have risen 44 per cent compared to the period 1990–2009. In addition, new conflicts entering the system after 2010 are up 150 per cent. As a result of these macro trends, the number of ongoing civil wars in the world is currently at an all time high.

Data Availability Statement

The data that support the findings of this study are openly available in “Replication Data for: Global Termination and Recurrence Macro Trends: A Follow-Up to Licklider and Dixon”, https://doi.org/10.7910/DVN/WB7AHD, Harvard Dataverse.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

1. Dixon examines the role of variables such as: regime type, demography, geography, economy, war type, military balance of power, war costs, and external interventions (military interventions and peacebuilding interventions) on civil war termination, duration, and recurrence.

2. This is according to the Uppsala Conflict Data Program (UCDP Conflict Encyclopedia); or more specifically, the ‘Number of Conflicts’ view at (https://ucdp.uu.se/?id=1) which shows state-based conflicts and non-state conflicts at the highest levels they have ever been.

3. There is only one Conflict ID per state for all govt incompatibility conflicts in UCDP.

4. From its outset, civil war termination and recurrence studies have tended to focus on conflicts that reach high levels of intensity relatively quickly. For instance, minor armed conflicts are not included in Sambanis (2004), Doyle and Sambanis (2000), Walter (2004), Quinn et al. (2007), David Mason et al. (2011), Mattes and Savun (2010), Walter (2004), Walter (2015), Gurses and Rost (2013), Glassmyer and Sambanis (2008), De Rouen and Sobek (2004).

5. Excluding minor conflicts resulted in dropping 87 of 194 Conflict IDs. Those 87 conflicts contain 222 dyadic episodes (out of 1178) and 152 conflict episodes (out of 460).

6. This was based on Appendix Table 1 in Powell and Thyne (2011).

7. The following timeline summarises information from the UCDP Conflict Encyclopedia entry for the conflict (https://ucdp.uu.se/conflict/288), as well as termination outcomes from Kreutz.

8. Thus the peace agreement did not terminate fighting for one year and is not counted as a termination.

9. Examining the endnotes in Licklider (1998) on this claim, he points to his 1995 article (Table A-1, pp. 688–689). Licklider had an N of 90 cases with roughly two-thirds terminating in the 50s, 60s and 70s, and one-third of the cases terminating in the 80s and early 90s.

10. Number of active dyads represented with a five year moving average of annual sums.

11. Note: Number of active conflicts represented with a five year moving average of annual sums.

12. The UCDP (and by extension Kreutz 2010) codes governmental conflicts in such a way as there can only be one civil war fought over control of the government in a country. If a governmental civil war is terminated and any new fighting group emerges in the future seeking control over the government (as opposed to seeking territorial autonomy or secession), this constitutes a conflict-level recurrence in UCDP even if none of the previous groups are involved.

13. Table 3 of Licklider (1995, p. 687) shows 14 peace agreements and 43 military victories, or roughly a 1 to 3 ratio.

14. We do not present the dyadic version of this graph as it is almost identical to the conflict version.

15. For example, the PAIC dataset (Political Agreements in Internal Conflicts) is a collection of hundreds of peace agreements and there is no requirement that the conflict be terminated. In 72 per cent of the agreements in PAIC, the conflict continues or recurs (Fontana and Neudorfer 2021).

16. The measure of recurrence used is: did the fighting stop for at least 5 years at the conflict-level.

17. The figure shows a loess smoother fit to annual observations of UCDP’s best estimate of one-sided violence committed by the government or the rebel group. In order to reduce the impact of extreme outliers, we dropped the 1994 Rwandan Genocide.

18. We caution that the estimate of government committed OSV following rebel victory after 20 years is based on a relatively small number of cases.

Additional information

Notes on contributors

Jason Quinn

Jason Quinn is Research Associate Professor of Political Science at the Kroc Institute for International Peace Studies at the University of Notre Dame. He is a researcher and a member of the leadership committee of the official verification mechanism in the 2016 Colombian Peace Agreement. Quinn’s scholarship focuses on civil war termination, recurrence, peace agreement design and implementation. He has provided research support in these areas to negotiation processes in Colombia, Myanmar, Nepal, Syria, Venezuela, South Sudan, and the Philippines. Quinn’s work has appeared in the British Journal of Political Science, International Studies Quarterly, Journal of Conflict Resolution, Journal of Peace Research, Negotiation Journal, Research and Politics, Defense and Peace Economics, International Studies Perspectives, Social Science Research, International Interactions, among others.

Matthew Hauenstein

Matthew Hauenstein is an AETL data scientist for the Lucy Family Institute for Data & Society. He holds a Ph.D. in Political Science from Florida State University with a specialisation in quantitative research methods. As a data scientist, he works with the Applied Analytics & Emerging Technology Lab. Before joining the Lucy Family Institute he was a postdoctoral scholar with the Kroc Institute for International Peace Studies at the University of Notre Dame.